Sunday 12 April 2020

Synthesizer Build part-27: QUADRATURE LFO Bergman edition.

A very quirky sinewave LFO with 4 outputs with a 90° phase shift to eachother. With extra waveform and an input for Frequency Modulation.

The quadrature LFO is something you won't find in many modular set-ups from the early days but it's a real little gem of a module to have in your set-up. It produces 4 sinewaves that are shifted in phase by 90° each. This can be used to create the 'Barber Pole' or 'Shepard Tone.' effect. "What's that?" I hear you ask. In short, it's an effect whereby it seems a tone is continously rising (or falling) without actually seeming to get any higher (or lower). See this Wikipedia Article for more on this effect. So that's one use. You could also feed four Voltage Controlled Panners with this LFO and have the sound go three-dimensionally around the room. Your imagination is the limit with this module which is why I was keen to experiment with this.

Eurorack 12V vs 15V.
I tested this circuit on a dual 12V powersupply and it works the same as on 15V. The maximum amplitude of the sinewaves stays the same too (+/- 7,5 Vpp). I think this is due to the zener diodes. They keep the amplitude constant. The maximum amplitude of the Trapezoïdal wave however will be lower at +/-10Vpp maximum. You can of course turn the amplitude of the waves down with the Output Level knob to, for instance, the Eurorack standard of +/- 5Vpp. No problem. So this is, in the way it works, a very Eurorack friendly module. For Eurorack you can cut the stripboard in half along the cuts in the copper strips, in the middle and fold it over. Then connect the top 3, strip J and the bottom ground copper strip together again with wire and you have a print small enough to fit behind a Eurorack panel.

How I came across the Trapezoïdal function:
I found this module with the schematic on the Yusynth website and set out to make a layout. Once again it turned out I made a faultless layout but it was with building it up that I made a mistake that turned out be an asset later on. This design uses two 5V Zener Diodes and somehow I had managed to stick one of them in the wrong hole on the stripboard so it was not connected correctly. This resulted in the output being Trapezoïdal instead of Sinusoïdal. Not knowing I had made this mistake, I posted about this on the Synth DIY Facebook page and people allerted me that it had to be the Zener Diodes that were at fault and they were right. After correcting the mistake everything worked fine but I then got remarks that I should include this waveform option in the final design of the module. So I did. I adapted the stripboard layout and added a switch to go between the two waveforms. If you want to build it as originally intended, without the switch, then just don't cut the copper between the two connection points for the switch on the stripboard, so the zener diode is connected again.
One more very important thing to note: YOU MUST USE POLYSTYRENE OR POLYESTER CAPS FOR THE TWO 10nF CAPACITORS!!  I've had several people build this module and then it didn't work because ceramic caps were used. They need to be high quality non-ceramic ones or it simply won't work.

ATTENTION: I have updated the layouts below in order to make them easier to read. I'm leaving the old layouts in the article for a while for those people who are building this project at the moment and used these as their guide. (They are fine btw).
I'm posting the new layouts below the old layouts and I'll the delete the old ones at a later date. So if you plan building this awesome LFO please use the new layouts.

Here's the layout. It's verified. I used it for my build. This is the wiring diagram. Beware the outputs are not in the right order. I added the correct order by numbering the jacks. All potmeters are viewed from the front.:


(Last revised: 12-April-2020. Corrected wiring orientation of dual gang potmeter and numbering of LEDs. 15-April-2020: Added mounting bolt and pin 5 and 10 warning text. 14-Nov.-2020 Added remark about capacitors to use.)

Stripboard only only. The four resistors to the top left and right of the LM13700 are 820 Ohms. It's a bit difficult to see on the layout. 
Beware that some stripboards are sold with 56 instead of 55 holes horizontally. The layout is 55 holes wide:


Bill of Materials:



NEW LAYOUTS:
Here are the new layouts, which should be a bit easier to read. I added more cuts to the right of the stripboard to make the copper strips a bit shorter and to prevent short circuits with anything touching the side of the board. Anyway, use these if you start building this project. I'm leaving the old ones up for now for reference and troubleshooting for people who already built this LFO.
These new layouts are a bit different compared to the old ones in that some cuts and components are positioned a bit differently. So don't mix new and old layouts together. Use one or the other.

Wiring diagram:
The current limiting resistor for LED D4 (bottom right) is a 12K and not a 1K2 like the other three. I did this because otherwise the white LED would be much brighter than the other ones. So it's not a mistake, it should be 12K for any white LED you use.


Stripboard only:
Red wirebridges are connections to +15V, Pink ones are connections to -15V, Green ones are connections to ground and the blue ones are internal connections between components.


Cuts and wirebridges as seen from the Component Side!!
As always, mark the cuts with a Sharpie or Edding pen and then put a pin through the marked holes and mark them again on the copper side. Then cut the strips at the marked holes with a 6 or 7mm sharp hand-held drill bit.


THINGS TO KNOW AND SCHEMATIC:
The LFO needs a few seconds to start up, when you first switch it on
. If the 'Rate' potmeter is turned all the way counter clockwise it won't start up at all. With Rate set to 1Hz it takes a good 20 seconds to start-up and before the LED's light up. So make sure it's not on minimum when you switch on, or you could have a long wait on your hands. Turn Rate all the way up and it'll start up almost immediately.
The normal sinewave output goes from -7.2V to +7.2V with a frequency of about one cycle per 30 seconds to about 140 Hz. You can set the lowest frequency you want available with the trimmer on the stripboard (T1). I had a warning from someone on YouTube not to set the trim pot all the way down because his went up in smoke. I had no such problems when testing this print but I just thought I'd mention it here. Anyway, he must have done something wrong because that shouldn't even be possible.
The Trapezoïdal wave has a higher amplitude and lower frequency so beware of that when you switch between them. It's maximum amplitude is -14,4 to +14,4 Volt and maximum frequency is 52Hz compared to the Sinewave's 140Hz. So you can roughly say the Trapezoïdal wave is double the amplitude and half the frequency of the Sinewave. That means both waves go high enough in frequency to be in the audio range so you can hear them. That opens up a wide range of uses for this LFO. One of my favourites is to have the CV control the Cut-Off frequencies of multiple filters and then feed the Quad LFO FM input with a ramp wave from an other LFO so you get a sweep. Sounds very cool!
The amplitude of the waveform(s) can be set by the dual gang 100K potmeter so you can set it at any level between zero and the maximum I just mentioned. The direction of the phase shifting between the 4 outputs can be turned around by the DPDT switch. You can see the direction it has by four 3mm LEDs on the panel. I didn't have an Orange LED so I used Red, Yellow, Green and White LEDs and the white one is quite a bit brighter than the rest so that's why it has a 12K resistor as current limiter and all the others have 1K2 resistors. This keeps the brightness perfectly in balance with the other three LEDs. One little drawback is that the LEDs only come on when the dual-gang output level potmeter is at the ten o'clock position or higher. At the lowest levels the LEDs don't light up. That's just how the circuit works.
Instead of the LM13700 you can also use the LM13600 and the TL074 can also be a TL084. From testing I also came to the conclusion that the 10nF caps don't necessarily need to be matched so closely. But remember they can NOT be ceramic capacitors. They don't work in this circuit. The two PNP transistors do need to be matched but you can use your multimeter's transistor tester for that and matched them to within 2 points of their hfe or amplification factor.

Here's the schematic I made the layout from:



I made a Falstad simulation of this circuit and it shows the start-up time really well. When you drag the 'RATE' potmeter to the left (faster rate) it will start up quicker. If you drag it the other way it will take quite a while.

Here's a picture of the finished panel:



Here are some screenshots from the oscilloscope:

Here you can see the beautiful and absolutely perfect sinewaves this LFO produces:



Here's the feature I added myself. Trapezoïdal waves. You can see the amplitudes differ a tiny bit from eachother but that's not a problem in normal use. It's the same with the Sinewaves btw.



Fast Fourier Transform (FFT) image of the sinewave at 11Hz. As you can see, it's a near perfect sinewave at this frequency. Almost no harmonic spikes to the right of the main wave. When the frequency gets higher the sinewave becomes a little less perfect but that's only visible on a oscilloscope's FFT readout, not with the naked eye:



Here's what happens when you add a Ramp wave from an LFO to the FM Input. You get a frequency sweep:




Finally a little video showing all the different aspects of this module.



Here's a video of the first test I did with the Quadrature LFO. The patch is set up as follows:
3 VCO's are feeding 3 filters with a squarewave. The Steiner, Korg and the ARP filters. The Steiner-Parker receives two sinewaves from the Quad-LFO, one on CV-1 and one on CV-2 which are 90° separated in phase. The FM input of the Quad-LFO is being fed by a 0 to 10V ramp wave from the other LFO. Each filter output goes in a different channel of the mixer where they are summed and the output then goes through the CaraOK effects unit set to preset 61 which is a chorus effect. From there it goes to the speakers. The "da-daa-dang da-daa-dang" drone you hear is produced by the waves from the Quad LFO. :)




Here's an other short sample of the Quad LFO. It's being fed by a slowly rising ramp wave and two of the outputs are going into the Steiner Parker filter. The ARP and the Wavefolder are also in the mix:



Just for my own record keeping, here is an image of the state of readiness my synthesizer is in now. I have 8 free power-buss connections left so plenty of room to build more modules in the near future.



Okay, that's it for this one. With special thanks to the folks over at the Synth DIY FaceBook Group for all their help in the initial troubleshoot.
As always, if you have any questions or comments please put them in the comments below. 

If you find these projects helpful and would like to support the website and its upkeep then you can buy me a Coffee. There's a button for that underneath the menu if you're on a PC or Mac. Or you can use this PayPal.Me link to donate directly. All donations go towards the website and projects. Thank you!

10 comments:

  1. hello !
    Is the diagram ok with eurorack +12V instead of 15V power supply ?

    ReplyDelete
    Replies
    1. Yes it will work fine on 12V and the amplitude of the sinewaves stays the same 7V wether you use 15V or 12V so no problems there.

      Delete
  2. Hi!
    Could you give an advice for solving the problem I have with this module?
    I'm not able to get 90 and 180 degree sides of wave shape symmetrical even if I have tried many thing I can do, like replacing all the ICs, zener diodes and trying several TRs that I matched by hands. 0 and 270 degree sides of wave shape is not also symmetrical like you attached above, but somehow I can go with it. However the other side, 90 and 180 degree coming from 7pin of U2 are asymmetrical and even there are some steps. Where do you think I have to check?

    Thanks in advance.

    ReplyDelete
    Replies
    1. I'm sorry you're having problems with it. I can't suggest a troubleshooting strategy for you, without seeing the stripboard but what I always do if I can't find the problem is simply build it up again. Take a fresh piece of stripboard and rebuild the module. Be precise and mark off the components you soldered in on a printed version of the layout. I wouldn't know what else to suggest.

      Delete
  3. Hi Eddy,
    I built the PCB project version of this and I am unable to get any part of the circuit to oscillate. I notice you mentioned this could happen if not using high quality matched timing capacitors. I matched my generic Tayda polyester caps to 0.1% with no luck. Could you by chance provide a link to the capacitors you used? I'm hoping I can solve my issues by replacing them.

    ReplyDelete
    Replies
    1. Polyester caps should be absolutely fine. Anything but ceramic caps should work. I used polystyrene caps but I still had those lying around. Bought them in the 90's.

      Delete
    2. I'm very late in saying this, but thank you for your reply! I totally forgot to check back.

      Delete
    3. No problem Wesley, I just hope you got the module working in the end.

      Delete
    4. Here I am over a year after my original post above - I locked myself in a room with this module and a pot of coffee and refused to leave until I got it working. It finally works! Used the wrong value for R19 and somehow repeatedly overlooked it. Caps had nothing to do with it. Go figure. Thanks for posting this page!

      Delete
    5. LOL, sometimes you can just stare yourself blind on a problem and overlook the obvious. Good show for sticking with it and getting it fixed!

      Delete

Note: comments are moderated and do not appear straightaway. Your first comment is not allowed to contain any links.