Synthesizer Build part-43: VACTROL VCF-1 by Skull & Circuits.

This is a bit of a weird one among filters and because it uses Vactrols it doesn't sound like any other filter. It has some percussive qualities too if you feed the CV IN with an LFO.

After not having built anything for two months I decided to get going again and this filter seemed like a nice one to try. It's a very simple design and thus easy to build. Because it's such a small stripboard and because it works on a dual 12V powersupply I made the layout more Eurorack friendly.  I have not tested this circuit on a dual 15V powersupply but I think it will work just as well on 15V.  

This project is more of an experimental one than one where the end product behaves in a way we expect from VCFs. What I mean is that the audio from this filter sounds different to all other filters and the CV IN behaves differently too. When I put an AD (envelope generator) signal on the CV IN, the resonance only comes up as the AD signal fades out, the opposite of how other filters behave and when I try to use a negative AD signal by putting it through the Dual Voltage Processor's Attenuverter, it doesn't give the expected result. Not even when I apply an offset voltage to the AD signal. So this is an experimenters dream. It works, but not like you expect. Just know that before you proceed to build it. ^____^

VACTROLS:

This is a Skull and Circuits design and you can find the original information by clicking here.

This is a 12dB LowPass filter, so a two pole instead of the usual four pole filters  This means the frequency rolloff won't be as steep as other (4 pole) filters. 

The Vactrols are something you'll need to make yourself. Well, you can buy them ready made but I advise to make them yourself because new ones can be hard to find and expensive and the ones from China will be fakes, I guarantee it. In the information link above is a video that shows you how to make your Vactrols step by step. I used two 5mm green LEDs (you need two Vactrols) and two LDRs no. 3P (or maybe 39) from a set with 10 pieces of 5 different types of LDR from eBay. They come in 5 little plastic bags and they are all numbered. I used the ones from the bag numbered 3P. I had used those before on the Lopass Gate and they react reasonable fast to light changes and go way down in resistance. The way your filter will turn out will be different from my filter. There are so many variables here that filters may turn out to be alike but never quite the same. This is mostly due of course to the Vactrols and the way you set the three trimmers. I tried two different types of LEDs in my Vactrols. First I tried bright white LEDs but with those I needed to turn most of the trimmers completely clockwise or counterclockwise to get the filter to behave correctly. The second time I used normal red LEDs (5mm) and they worked better. That's what I used in this build. 

Later on I bought the kit from Skull & Circuits and in that kit I used Vactrols made from green LEDs which I found work even better than red ones. (The kit built filter sounds the same as the stripboard one.)

Now I could use the trimmers more accurately to get the right sound. So I advise to use either red or green 5mm LEDs for you Vactrols. The LEDs in the filter itself, mounted on the front panel, do not shine very bright in my module. I used two 3mm yellow LEDs and they only come on when the trim-potmeters are set to extreme positions. I can get them to shine brightly when I turn the trimmers a certain way but then the filter doesn't work properly. Again this is one of those things that will be different with your filter because of component tolerances and differences in the Vactrols etc.

Just out of interest, here are some measurements I took of my DIY Vactrol (out of circuit) with red LED:

OFF resistance = >200MOhm.  ON resistance with a +/-5Vpp signal = about 10K, with a 0/10Vpp signal = about 5K. This is with a 1K resistor in series with the 5mm red LED of the Vactrol.

Of course these values differ when the Vactrol is in circuit because there you can set offset voltage etc. which alters the resistance. With the green LEDs, in the kit I built later on, I could get even lower resistance down to 200 Ohm!

NOTE: This module pairs extremely well with the Voltage Processor and the Lopass Gate module. If you create a beat with an LFO into the CV IN of the filter but first put the CV through the voltage processor you can accurately control the sound and if you then send the audio through the Lopass Gate you can turn your beat into a sort of galloping beat. Connect a signal from the same LFO used for the VCF to CV1 of the Lopass Gate and use a faster LFO signal on CV 2 of the Lopass Gate. Used in this way the Vactrol VCF can be used to make some cool Techno kickdrum sounds.

LAYOUTS:

Here is the layout I made for this filter. As always the layout is verified. You can see it's small enough for a Eurorack module and I incorporated a Eurorack power connector to make life easier for those of you using that format. I myself didn't use the connector in my build but just soldered the power cable straight to the print like I always do. All potmeters viewed from the front.

IMPORTANT:

Stripboard only view below. Note that the copper strip between pins 3 and 12 of the IC is NOT CUT! Both pins are connected to eachother underneath the chip. Also do not forget the cut at hole B-29 otherwise, when you close the Cutoff potmeter, you'll create a short circuit between +12V and ground through the wiper of the potmeter which will produce smoke, I guarantee it ^^ :

Strangely enough I found that I needed to reverse the wiring of the Cutoff potmeter for it to work the right way around. I left it the way it is presented on the schematic, in the layout so you can decide for yourself whether you need to do that or not. 

Bill of materials:

TRIMMING THE CIRCUIT:

Trimming this circuit is a matter of trial and error. You need to set the trimmers in such a way that you get a good deal of self oscillation when the Resonance potmeter is set almost completely open and you need to get the right range for the Cutoff potmeter. I can not give you a procedure to help you do this. You will have to figure it out yourself but it's pretty straight forward. Just put a squarewave or rampwave signal on the audio input and connect an LFO signal (+/-5Vpp) to the CV input and then connect the output to your VCA. Changing R7 from a 47K resistor to a 100K trimmer helped a lot in getting this filter trimmed. It was suggested in the text of the original Skull and Circuits article about this VCF and I also saw they implemented it on the PCB so that's why I put in a trimmer for R7 in the layout. It might be a good thing to use multiturn trimmers instead of the single turn ones I use in the layout. I used normal trimmers on my board but it can be fiddly to set these correctly. I might replace mine with multiturn ones too. You might think it would be a good idea to put an attenuation potmeter on the audio input but I tried that using the passive attenuator on my mixer but it didn't help at all. The filter only worked if I had the potmeter fully open so there's no need for an audio input level potmeter. As you can see in the schematic, the audio is attenuated by a factor of almost 18 times by the 1K and 56 Ohm resistor voltage divider right at the audio input. Then it is boosted up again in the output opamp.

The whole upper portion of the schematic dealing with the Cutoff and CV input goes straight to the LEDs of the Vactrols so you can see that he whole filter hangs on these Vactrols. The Offset and Range trimmers control the Vactrol LEDs and determin when the LEDs turn on and how sensitive they are.

The LEDs over the opamp go on when the signal is clipping. You should try and trim the potmeters so these LEDs turn just about on when the filter is in full resonance and cut off is 3 quarters closed. So when the filter is really working its hardest these LEDs flicker in time with the CV input voltage. These LEDs should not be burning at full brightness in normal use!!

Schematic drawing:

Here are some pictures from the build proces. In these pictures the third trimmer is not yet put in. I tested it first with R7 as a normal 47K resistor. I later changed it for a 100K trimmer.

The stripboard:

Stripboard mounted behind the panel not yet wired up:

Video of the first test. You can see here that connecting an AD (or Envelope) signal to CV IN results in the opposite reaction to what normally happens. Using a negative or inverted AD signal doesn't fix this issue. Not even with a positive offset voltage applied. That just proves that this is a very quircky filter and all the more interesting for it ^__^  I've heard better results than mine from other people who built this VCF, so it depends a lot on how your vactrols turn out.

If you have any questions or remarks you can put them in the comments below or go to the Facebook Page for this website where we have an awesome little community of cool DIY synthesizer enthusiasts that are more than willing to help you out with any problems you might encounter.

Synthesizer Build part-11: ECHO and SOUND FX UNIT and LINE OUT.

100 different sound effects combined with a line-out and head-phones connection for the DIY synthesizer.

Here's a little item I found on eBay and thought it would work great in the synthesizer; and luckily I was right. It is a fantastic asset and makes the synth sound really professional and full. 

A word of warning though, this is not a beginners project. You need to know your electronics to follow my plans. This unit also needs two voltages to run on. The effects unit runs on +5V and the stripboard runs on +/-15V or +/-12V. So you must make sure you have those voltages available in your modular set-up or add a +5V regulator to the print. I recently added a Dry/Wet control so you can dial in as much of the effect as you want in your sound. This makes this unit really useable and perfect for modular synths. (see further down the article)

The unit goes under different names but mostly as the Cara OK ("Karaoke" get it?) or DSP 5V Red Digital Stereo Mixer Reverberation Karaoke Reverberation Board Module (if you copy and paste that into the search bar on eBay, you'll find it). The prices vary from $15 to about $30 us. (DSP stands for Digital Sound Processor)
This is the cheapest listing I could find on eBay: click here.
It offers 100 presets with reverb, echo and even chorus, phaser, flanger, phase shift and reversal effects and combinations of them together. There's a rotary encoder with which you choose the preset of your choise and then you just press to confirm and engage the effect. It's a favourite with many synth builders I noticed.
I made a special panel for it and combined a 'Line-Out' control and bypass switch option with the module. Here is the schematic or wiring diagram to make this unit part of the synthesizer. This was one of my early projects and at first I didn't have dry/wet control in there so this is the version just with a switch between normal line out or line out with effects:

I had to put extra attenuation on the bypassed (normal) signal because it was much louder than the output from the Echo Module, but that was easily fixed as you can see in the diagram above. The panel that I built into my synthesizer also includes a headphones out stereo jack which is not included in this article but it's just an output jack soldered straight to the negative poles of the Left and Right output Electrolytic Capacitors. Weirdly enough, if you listen through the headphones, the echo module is much louder than the normal line out. This is probably some impedance matching issue but it doesn't bother me. It's easy enough to turn a volume knob so I'm not bothered.

The Stripboard can work on +/-12V also. It doesn't matter if you use 15 or 12V. The CaraOK effects unit must have it's own +5V powersupply! You can use a 7805 voltage regulator and connect it to the 15 or 12V of the stripboard to get 5V for the effects unit. This is not further described in this article. I'm assuming you have the knowledge to make powersupplies. If not, check out my article on this subject here or see the datasheet for the 7805 voltage regulator.

Here's the stripboard layout (older version). This one has a 3 pole switch but that's not necessary. Just connect the wires on the left hand side of the switch together; the red line coming from the 'normal level pot', the blue line from pin 1 of the chip and the effects unit input. They can all be connected together and then you just need a 2 pole dual throw switch (DPDT).
(All potmeters viewed from the front)

Please note: The GROUND connection of the Effects Unit must be connected to the GND of the stripboard. Everything must share the same ground. I myself made a central grounding point with a solder eye connected to one of the M-3 bolts holding the Effects Unit to the panel. To put it in a simple way: the ground of the 5V powersupply for the Effects Unit must be connected to the ground of the +/-15 or 12V powersupply of the stripboard. The Grounds of the Line Out cables are also connected to this common ground. You can tap that off from the 4th strip above the Left Line Out connection. (2nd strip below the chip).

Print only:

(Last revised: 14-Aug-2020: Corrected mistake with negative voltage supply to the TL074.)

I used a 3 pole double throw toggle switch (ON-ON type) to be able to switch the synthesizer between normal output and output through the effects unit. This is a bit of over-kill because you can just as well connect the inputs together (part S1-A of the switch) and then use a 2 pole switch to switch between the outputs. Better still. I describe down below how you can put in a dual gang potmeter instead of a switch and so have a DRY / WET control. 

I'm going to adapt this article soon and make a new layout to put in this Dry/Wet control permanently. 

[EDIT] March 2021 DRY/WET CONTROL:

I wanted a Dry/Wet control on this effects unit for a while now and I wanted to install it without having to rebuild the whole module.

I changed the 3 pole toggle switch for a dual gang potmeter and I connected the inputs, normally connected to part S1A of the switch, all together. If you do this make sure you keep to the right order with the wires. Best to make a few pictures of the switch connections first before you solder in the dual potmeter, that is, if you built this module already.

Below here is a schematic. You can see the switch has been replaced by a stereo potmeter. This MUST be a linear type potmeter!!

There is a half drop in volume at the half way stage of the potmeter because we have 50K of resistance in our signal path there but other than that it functions fine! It's just a matter of turning up the input or output levels to get it where you want it. You must make sure the output audio goes into a very high impedance input, like a HiFi amplifier because if you pull even a tiny bit of current from this circuit it won't behave normally anymore. But it works fine on audio amplifiers, I guarantee it.

The Dry/Wet control really makes a world of difference! Now you can set it to reverb and then precisely dial in the effect to where you like it. It's fantastic sounding!

Here is a picture of the panel with the Dry/Wet control where the switch used to be:

I made a little demo video. You can hear the drop in volume at the midway point of the Dry/Wet control. (I know it's wired backwards LOL :)  But being able to dial in the effect makes a world of difference and the volume can easily be crancked up by the level controls.

Continuing the original text:

I've put in 4 buffer stages, using the TL074, for the input, normal output and FX Unit outputs Left and Right channels and I gave the latter two adjustable gain by means of two 50K potmeters in the feedback loops of the opamps; one for each channel. The gain is adjustable from 2 to 5.3 times. You can increase that by using 100K potmeters instead of 50K ones. That would give a maximum gain of 8.6 times.

When I first tested this unit I noticed I was receiving an FM broadcasting signal through the effects unit. (There's an FM Broadcast transmitter and antenna on a flat 100 meters from my location). So I took a ferrite ring and wound the audio input wire around the ferrite ring about ten times. I also put ferrite beads in the 5V power-supply line to the effects unit and to the print with the opamps on it. This solved the problem. One more little thing: beware of the little crystal X1 near one of the screw holes on the circuit board. It is rather flimsy and fragile. Take care not to damage it.

The Cara OK is a really versatile unit with lots of really cool sounding effects. Here's an overview of the possibilities it offers. I myself printed a small version of this list out, laminated it and stuck it at the bottom of the panel I made for it. Handy to have around I thought :) :

This picture shows all the connections to the circuitboard:

It's small so it won't take up too much space. Beware that it needs just +5V for power supply. Luckily in my synth build I made a power-supply that delivers dual 5, 12 and 15 Volts so I can feed it right from there. I can really recommend you picking this up and trying it in your build project. It will add a lot of options and is a very useful addition to the filters and its output is in stereo! The sound quality is just great so no problems there. The only thing is the difference in volume I mentioned earlier but that is easily fixed. You can use opamps buffers with it, like I did, but it's not absolutely necessary. I did without them at first but then installed output buffers with variable gain as I mentioned before.
The audio response of this module is so good that it even lets through the ultra low frequencies the Korg MS20 filter produces (see next article) and that can go as low as 10Hz. You can really see the speaker cones move bigtime!
Before I installed the 3-pole toggle switch I had a single pole and a double pole switch side by side to switch between FX-unit and normal line out. So after installing that 3-pole switch I had a hole in the panel left over. I mounted a 3,5mm stereo output jack in that hole as a connection for head-phones. The output jack is connected straight to the audio output on the stripboard. One thing I noticed with this arragement is that the normal line-out through the head-phones, sounds a lot quieter than when the effects unit is switched on. That's probably due to a difference in output impedance because we're effectively putting an 8 Ohm resistance between Line-Out and Ground in the form of the head-phone speakers. This doesn't occur when I listen to it on the normal audio amplifier, at least not if the head-phones are not plugged in. It would be a good idea to build a little head-phone amplifier for this purpose.
Here is a picture of the finished module in my synthesizer:

Here's a picture of what's behind the panel. Now you understand why this is not a beginners project ^_____^  You can see the yellow Ferrite ring with the black 'audio in' wire wound around it above the blue circuitboard and there's also one on the stripboard. I advise you to include these in the power supply line and audio in line. In red you can see the 3 pole toggle switch. This panel works really well like this.

Here's the Line Out Panel I made on the back of my synth, with two gold plated RCA outputs and a 6,3mm (1/4") Stereo Output Jack, which is connected straight to the RCA left and right outputs.

ONE LAST REMARK:

I have had feedback that some chips on the CaraOK board can run hot when run for a long time and even fail after a few years so it is advisable to add an ON/OFF switch to the module so you can turn it off when it is not needed thereby preventing the chips from being on needlessly for a long time.

(see comments below)

Okay, that's the 11th module I put in the synthesizer. We're nearly there. I have just room enough left for one more module and that has got to be the Korg MS20 filter. But I'm waiting for some supplies from China before I can build it. (Circuit boards for one, coz I'm fresh out at the mo.)

Okay here's an excellent video by Juanito Moore that shows you how you can circuit-bend this device and make it voltage controllable. Click here

Right, that concludes this article. Thanks for stopping by and while you're here, why not leave me a comment or if you have any questions put those in the comments too and I'll get back to you asap.

CZE T251 FM Broadcast transmitter REVIEW

This is an overlook and general impressions review of the Chinese made CZE T251 FM Broadcast transmitter. I ordered mine in 2017 out of China from eBay seller 'Thanksbuyer-hobby' for $219,- including shipping. It took 11 days to arrive which is fast!

Update 1st of November 2020.  Before we start: I haven't been working with FM transmitters in the last 2 years but I had a look lately and it seems they have gone up in price quite a bit. I could also find only one listing of this particular transmitter on eBay. Just so you know, they are getting rarer and are probably being replaced by newer models but I can not give you any opinion on those. As I said I no longer occupy myself with transmitters. It's all synthesizers now =) Therefore the comments for this article are now also closed. The new GD-20155-ALL 15 Watt FM transmitter using the BH1414 chip looks like a good one to buy though.

BUT KEEP IN MIND: Chinese equipment is okay as long as it works but when it goes wrong and you need to refer to customer service you're in for some big disappointments.

Normal operation:
You switch on by pressing the on/off button at the back. Then you get either the standby screen or the transmit screen, depending on how the transmitter was set when it was last switched off. If you press the front dial for more than one second the transmitter goes into sleep mode and won't transmit. The button will light up red. A simple short press of the front dial button awakes it and powers it up. The button now lights up blue.
Here are some pictures to illustrate this:

In transmit mode the display shows, from top left to right: the frequency it's transmitting on, the audio volume (goes from 0 to 79, you can see in the picture above the audio is set to the maximum 79) and whether it's stereo or mono (one or two speakers displayed) and the microphone level (also from 0 to 79, M00 means microphone level is set at minimum). The row underneath shows the RF power output, the SWR reading and the temperature of the power transistor. When transmitting, turning the front dial regulates the output audio volume of the transmission. You can access the menu while transmitting and change settings without the transmission being disturbed but if you change the RF frequency, the transmission will stop.

The Menu consists of the following items:
1 - Frequency set
2 - Audio Volume
3 - Microphone Volume
4 - RF Power set
5 - Stereo or Mono
6 - Temperature Alarm set
7 - SWR Alarm set
8 - Mute
9 - Exit Menu
If you turn the dial clockwise you go through the menu from item 1 to 9. If you turn counter-clockwise you go through the menu from item 9 to 1.

Things to be aware of:
First thing to be aware of when you buy this transmitter without a power supply is that it needs a 2.5mm inner diameter/5.5mm outer diameter DC barrel connector. Not the standard 2.1/5.5mm ones that are used with, for instance, the CZE/7C. The correct procedure to switch this unit off is to first press the button at the front to put it in sleep mode. If the fan is running when you switch off like this, it will keep running for a short while (about 20 seconds, give or take) and then, when the fan stops running, you can switch the unit off by pressing the on/off button at the back. If you switch off following this procedure, the unit will always start up in sleep mode. This prevents it from immediately transmitting when switched on. A simple short press of the front dial button will then start the transmission. The RF power will start rising smoothly to the set output power and will reach that power level within a maximum of approximately 10 seconds.
Here are two pictures of the backside:

Also something to be aware of is that this unit has as standard an N-type RF output socket (female). I screwed an adapter plug (N-male to SO239) onto the antenna output socket to change it into an 'SO-239' because I only use 'PL-259' plugs on my coax.
If you look closely at the top two pictures you can see, I made two little extensions (from some Ø15mm copper tubing) for the front feet so the transmitter stands a little higher. The air intake is underneath and at the front so it's best to put the transmitter a bit higher on its feet to give it some space to breathe in air. If you don't do this the front panel almost touches the surface with the main body clearance being 4.5mm. On YouTube I've seen a few other users of this unit do the same thing. You don't HAVE to do this as long as the unit stands on a smooth flat surface.
Before owning this transmitter I had used the CZE-7C, 7 Watt transmitter and I must say I liked it a lot. Even-though many reviews complain about the sound quality of the CZE 7C, on my car radio it sounded pretty good but the T251 sounds even better! Anyway, it's good to have a CZE 7C handy. In my case I use it to test the bandwidth of my antenna and of Low Pass filters I build etc. So I sort of use it as a signal generator, because you can change frequency real easy whilst transmitting. That is something the T251 will not do! If you go into the menu to change the frequency, the RF stage switches off. Btw, I noticed that if you put the T251 in sleep mode after it has been transmitting it still transmits a very weak signal. It can't be more than a few micro Watts but I can still hear the music buried in noise on my radio. Probably the FM chip that still gets power from electrolytic-capacitors that haven't discharged yet.
The FM chip:
I used to think this transmitter was equipped with the BH1414K FM-chip but I inquired with the seller and they informed me that it has the BH1415F chip inside. This was a disappointment for me, because for the price I thought it would have the much better BH1414K, but still, it does a pretty good job. Of course it's not only the chip itself that is important. It's also the circuitry around it, and there is a distinct difference between the signal from this transmitter and the signal from the cheaper ones which also use the BH1415F like the CZE 7C. That's mostly due to the end-stage with it's filters and power transistor. The RD30HVF1 mosfet is an excellent and robust transistor for the end stage. However, it is noted in the comments below that the BH1415F chip can cause some overload issues that result in stations near to the transmitting frequency being overpowered by the CZE's signal for about a maximum of half a mile around the transmitter location. So be aware of that. It's not perfect (obviously).
Frequency stability is rock solid and stays well within the advertised maximum deviation of 10Hz.
Of course you can analyse the signal from this transmitter until you're blue in the face and it's never going to be as good as a professional Broadcast transmitter but that is reflected in the price.
Here is a picture of the output sinewave at 5 Watt RF power on an oscilloscope. As you can see a nice clean signal.

Pre-emphasis delay is 50µSec and can not be changed in the menu. So this transmitter is not aimed at the American market.  You will just have to contact the seller and ask if they have units with 75µSec pre-emphasis delay if you live in the United States. USA Stock is regularly promoted from different sellers on eBay. (I had a link here to some USA stock on eBay but the listing has ended.)
This transmitter has an RS232 connection at the back, underneath the N-Connector but don't think you can control this unit with your PC. That connection is only for debugging and firmware updates.
Don't assume either that the extra 18 Watts of power compared to the CZE 7C is going to increase your range significantly with the same antenna setup. It's not. You want bigger range? Put your antenna up higher. This set gives me just a little bit more range than the CZE 7C with the same open dipole antenna that I use and the signal is a bit stronger within the normal range but it's only a very small improvement. Believe me, like with an audio system, the quality of the speakers is everything, so it is with antennas and especially their height. (I address the point of 'range vs power' at the bottom of this review also.) As a rule, to double your range with the same antenna set-up, you need a 10 fold increase in output power!!

A closer look at the signal:
Alas I don't have expensive spectrum analyzers etc so here's a few screenshots of the signal from my Software Defined Radio (SDR) on my computer (click on the picture to get full screen view):

Stereo transmit signal at 25 Watts:

This is the signal from a professional local FM radio station whose transmitter and antenna are located just 100 meters from my house. Compare this signal to the previous picture of the T251. Pretty similar right?

Btw, if you look at the signal to noise ratio (SNR) you can see that it is a bit lower on the T251 than on the professional transmitter but that changes with the type of music. If there's a low volume bit in the music the SNR on the T251 can go as high as 37dB. Btw, that is SNR measured with this SDR receiver. That's not exactly a precision instrument. The real Signal to noise ratio is advertised as being equal to or above 70dB. And in fact, CubicSDR software indicates a SNR between 65 and 70dB so that is excellent. These numbers I mentioned are for the stereo transmissions. When you transmit in mono, the signal to noise ratio will be even higher and the mono transmissions also have a longer range than the stereo transmissions.

Mono signal:

Stereo signal but no audio input:

Signal comparison between the CZE-7C (7 Watt transmitter) and the CZE-T251 with the same song playing, both in stereo and both audio volumes and output power set to the same level. Note the difference in deviation beyond the 200kHz bandwidth and the more slender signal of the T251:

This is the CZE-7C. You can see in the waterfall display that the audio bleeds over the 200kHz bandwidth limit, causing disruption on the adjacent frequencies. You can also see that the signal is quite wide:

This is the CZE-T251, here you can see the audio stays nice and tight within the 200kHz bandwidth limit, like a good transmitter should do. The signal is also much more slender than that of the 7C:

Here are some screenshots done with "CubicSDR" software:

Signal at 25Watt without modulation. You can see the 19kHz stereo pilot-tone and more in the upper right audio spectrum graph:

Signal with normal music. Again note the audio graph in the upper right-hand corner:

This is an illustration of how the audio signal produced by an FM stereo transmitter is build-up and how it occupies the frequency spectrum:

You can find more on the technicalities of FM Broadcasting on Wikipedia.

More observations about the T251:
The transmitter is very well built. It's all thick Aluminium. There's no plastic on it anywhere except for the on off switch at the back. Even the dial knob at the front is made from Aluminium. It is a lot smaller in size then I expected from the pictures on eBay though. Be prepared for that. The size is: front-panel width: 173mm (6.81") x front-panel height: 58mm (2.28") x depth: 210mm (8.27"). Main-body width: 167mm (6.57"). Main-body height: 53mm (2.09")
The power amplifier part (end stage) is mounted to a big heat sink with ribs that go all the way to the side of the case and with a fan attached to one side which blows air over it. This is more than adequate to keep the temperature down. (More on the fan in the item below.)
The transmitter is equipped with a temperature alarm. If the temperature gets higher than the alarm setting the unit will stop transmitting and an alarm will sound.  I have set mine to 50°C. The power transistor can easily take 120°C according to the spec sheet but the fan keeps the transistor at a maximum of 39°C so if the temperature rises above that, then it's obvious that there's something wrong and in that case the sooner the unit switches off, the better.
The unit also has a build in SWR meter with a programmable SWR alarm setting. If the alarm is triggered the unit stops transmitting immediately and an alarm will sound continuously until you switch the unit off. If you then switch on again the RF power output has automatically changed to 1 Watt. This is obviously a safety measure in case there's still something wrong with your antenna system. If all is well, you can go into the menu and set the power to your desired value. The SWR meter is a bit optimistic. It says 1.0 when it is really 1.1 but that's no problem in normal use.
The power transistor used is the RD30HVF1 mosfet rated at 30 Watts. It has a typical efficiency of 60% (meaning 60% of the energy put in is transferred to RF power, the rest to heat) and can take an SWR mismatch of 20 to 1 without being destroyed (!!!) The temperature sensor is mounted near the mosfet and is of the LM35 centigrade type. Both the sensor and the power transistor are generously covered in a heatsink compound as you can see in the pictures in the link below. In use with a voltage of 12.55 Volts I draw about 3.37 Amps at the full 25 Watt RF-power setting, so that is 42.29 Watts. (this is measured without the fan running) So that results in an efficiency of 59,11%, almost exactly the rated efficiency of the power transistor which was, as I mentioned earlier, 60%.
The chip that controls the display and remembers the settings is the 12C5A.
You can set the input audio volume electronically from 0 to 79 and the same for microphone input. Beware that you need a dynamic microphone and NOT an electret type microphone!! The microphone input is a mono channel (6.3mm jack plug needed). But using a microphone directly connected to the transmitter isn't very practical, unless you're transmitting at a public event or something. Otherwise I would use a microphone connected to an audio mixer and connect the mixer to the audio input of the T251. Audio input sensitivity is a bit less than on the CZE 7C. The input level is rated at ≤-15 dBV. The frequency response is: 50Hz to 15kHz (3dB). Distortion is 0.2% and the channel separation is 45dB (The channel separation of the average tuner/receiver is about 35dB so the transmitter is well above that.) I use a Philips MP3 player and I have its volume set all the way up and the same on the transmitter (volume to 79) and that get's the music exactly to the level of other stations. But it's better to use a pre-amplifier for the music, like a headphone amplifier or the line-out from an audio mixer to give yourself some leeway. I pre-record my programs and then run it through Adobe Audition which has a very good compressor plug-in by iZotope with a 'Broadcast' preset. That's what I use to render the radio show out and then transfer it to my MP3 player for broadcast. This works very well and the music quality is great. This unit has very tight bass and crisp highs. The sound of the 7 Watt CZE-7C is muddier in the lows and just doesn't reach the quality of the T251. Audio input is in the form of RCA connectors at the back of the unit. In the transmitter menu you can choose between stereo and mono transmission and it is my experience that the mono signal gives you a much bigger range.

More on the cooling fan:
The cooling fan kicks in automatically when the temperature reaches 39°C (102 F) and switches off again when the temperature drops to 35°C (95 F). It also starts running if you enter the menu to change settings. I guess this is done as a quick test to see if the fan is working. The fan produces no extra noise in the audio but it does produce a loud noise of itself so you don't want this transmitter near your microphone. Your studio must be in an other room than the transmitter or you'll hear the fan running in the background. You could of course replace the fan by a more silent one. It's not difficult to remove. The dimensions of the fan are: 40mm high x 40mm wide and 15mm thick. The hole spacing is 32mm. The fan get's 11,0 Volts when the transmitter is fed with 12,0 Volts. The fan's power cable has a 3 pin connector on it of which only the outer two are used. The middle pin is not connected. So if you get a replacement fan with a 2 pin connector you could just cut it off and solder the wires directly to the wires of the connector in the transmitter. The fan draws a current of 118 milliAmps at 11 Volts.
Fan Replacement Test:
I ordered some cheap silent fans from eBay and ran some tests. The first fan I tested was 40x40x10mm 12V 0.08 Amp. and this was totally not up to the job. It could not get the temperature down and kept running all the time. The second one I tested had a thickness of 15mm and a current draw of 0.1 Amp. It could get the temperature down enough to make it switch off again but only for the first 15 minutes. After that the heatsink was fully warmed up and too warm for this fan and it couldn't get the temperature down to where the fan would switch off again. So I reinstalled the old fan. So if you plan on replacing the fan you will need a type that is 40x40x15mm and with a minimum current draw of 150 milliAmps at 12 Volts. This is a picture of the fans I tested and which failed the test:

Here's a picture I took after the test, with the original fan again connected but not yet screwed into place. You can see the heat-sink arrangement and the 3 pin connector for the fan:

I couldn't actually find a good silent fan that was up to the job. 

Range vs Power:
With the T251 set to stereo, the full 25 Watts output power and a dipole 7 meters off the ground and in free air but not clearing the rooftops (see picture of antenna below) and an SWR of 1.1 to 1, I get a range of 10 kilometers (6¼ Miles) in a small-town setting without high-rise buildings. (Nothing higher than 5 stories to block the signal). That is a 10 kilometer range in which the audio is noise free received on a car radio driving away from the transmitter location (receive antenna on the rear of the car roof, so the ground-plane formed by the car roof is pointing away from the signal)! The range in which the signal can still be received is about 15 to 20 kilometers (9 to 12½ Miles) depending on terrain and obstructions. If you put the antenna high enough to clear the rooftops of most buildings near you, that can easily double! I also did a test in which I set the transmitter to an output power of 1 Watt and tried what range I had. It was unbelievable. I still had a range of about 5 kilometers (3 Miles). Sure the signal broke up much faster behind obstructions like buildings and bridges but still. 5 Kilometers is not bad. As I mentioned I set the T251 to 1 Watt, but measured on my Diamond SX600 SWR/Power meter an output power of 1.75 Watt. But seeing as it has to go through 7 meters of RG213 and 2 meters of RG58 I recon the Effective Radiated Power (E.R.P.) couldn't have been more than 500mWatts. That's impressive and shows once more that the most important thing about a transmitter set-up is the height and placement (and build quality) of the antenna!! Height is everything!

Here's a picture of my home built dipole antenna:

If you want to have a look at the Printed Circuit Boards inside the transmitter, then I refer you to the link below to my Flickr page where you'll find an album with some closeups of the PCB's:
https://www.flickr.com/photos/ededitz/albums/72157679661456110

Troubleshooting:
Should you encounter any problems when operating this unit then there is an option you can try. If you switch the unit on with the switch at the back and at the same time keep the front dial button pressed in, the unit switches on into Self Test Mode. Here you can automatically check certain functions of the transmitter. Simply turn the dial to select the parameter you want to test and press to confirm. The unit then tests itself to see if it functions correctly. I personally have not tried this option yet because my unit functions perfectly fine and I am not comfortable doing this procedure when it's not necessary so I can't tell you exactly what it does when you select this option but just know that it is available should you need it.

This transmitter has one weird quirk. If you turn the front dial button whilst the unit is in sleep mode, with the button lit up red, or even if the unit is totally switched off, the temperature alarm can go off if you switch it on again! If this happens simply press the front dial button twice or switch off and on and you're good to go. A minor thing. Just don't play with the knob when it's switched off.

I've been getting some feedback that the output power is not really 25 Watt but more like 21 Watt. I myself measured this too on my SX600 SWR/Power meter but I thought it was my meter that was at fault; but I'm hearing from more people that they get the same measuring results. Now 4 Watt isn't a lot when you're already outputting 21 Watts so I shouldn't let it influence you if you're thinking about buying one of these units but I want this to be as honest a review as possible so I thought I would just mention it here.

On that note there is one more thing I must mention. A very small percentage of these units can fail and give a reading of zero output power and the SWR indicates 9.9 even though they go through quality control and are tested for a period of time before being sold. If that happens your only option is to contact the seller and try and get a new unit or a refund. The chances of this happening are extremely small but they are not 0%. Like anything you order from China there is a small gamble involved but usually the seller is very quick to replace the unit for you. It has happened to one person commenting below so you should be aware of this. But it is the first mention I received in the 2 years that this article is now online.

Conclusion:
The unit I bought and tested performed very well. Certainly not a 'toy' as some people on YouTube like to say when they compare this to a $5000.- transmitter. That's like comparing apples to pears.
It simply is NOT a professional FM Broadcast transmitter but listening to it on your car radio no one is going to notice that. Stations that are 0.2 MHz or more away from your frequency won't be interfered with by your signal. And you don't need to be afraid of harmonic signals either. The Butterworth Low Pass Filter in the end stage suppresses that very effectively. But do be aware that around the direct vicinity of the transmitter location, stations can be surpressed by this unit. This is however only within a few hundred meters of the transmitters location and it's not always the case. It depends on the local situation of course and also on the quality of your antenna. Professional broadcast people curse this transmitter, and ones like it, because they say it creates interference on other frequencies. I can not see how it can do that with a lowpass filter behind the powertransistor but then again I am not an expert, just a consumer who reports his own experience with the unit for others to make of what they wish.

The sound quality is amazing. As I mentioned before, the cheaper CZE 7C's sound is phatter and is muddier in the bass region than the T251. This unit sounds very crisp and clear with deep tight bass and crisp highs. Really excellent I thought. Of course you must keep in mind that the perceived sound quality is first of all dependent on the quality of your receiver and audio system. Make sure to check whether your transmitter produces interference on other frequencies before you start using it. Some of these could have fake power transistors in them, although mine was fine.
If you want to buy a Chinese FM transmitter, I would get this one (or the cheaper 15 Watt CZE-15B which does have the BH1414K chip inside and is PC-controllable.)
I would stay away from the 60 to 80 dollar 15 Watt ones like the NIOrfNIO or the ST-15B. They are cheaper for a reason. Their quality is not terrible but comparable with the CZE 7C I mentioned above. But certainly stay away from the cheap FM kits. They use the KT0803L chip which is the worst FM chip you can get. Their signal is a mess, you'll black out the TVs in a full block radius. Spend a few extra dollars and buy better quality.

Here is a link to the USER MANUAL for the CZE T251 in PDF form:
Click here for User Manual PDF

Here is a link to the datasheet (PDF) of the RD30HVF1 Mosfet power transistor:
https://www.alldatasheet.com/datasheet-pdf/download/170369/MITSUBISHI/RD30HVF1.html

The comments below this article are proving to be a great resource. Reading them will give you an idea of other peoples experience with this unit and how it compares with other ones. Unfortunately I had to close the comment section for this article because it became too long and also because I've been out of the FM transmitter hobby for so long now that I don't feel qualified to answer any more questions about it. Sorry. But read the comments below and you'll probably find an answer to your question.

Disclaimer: The author does not accept any responsibility for actions and or alterations to equipment undertaken by anyone after reading this blog post. Anything you do, you carry the responsibility for.
Operating an FM broadcast transmitter without a licence is illegal in most countries, however, owning a transmitter but not using it as such is usually not illegal. Please check the laws of the country you live in to make sure you don't break any laws.

-- SORRY. COMMENTS CLOSED! --