Wednesday 18 September 2019

Synthesizer Build part-1: SYMMETRICAL POWER SUPPLY (1,5A) with independent voltage adjustment.

The essential building block for our DIY synth. with outputs for +/- 15V, 12V and 5V at 1.2 Ampères.

Please scroll half way down the article to find the second version (which is modular in set-up) with stripboard layouts

I have set myself the task of building my own synthesizer. At the moment I'm slowly buying in the components I'm gonna need, as my budget allows (and that ain't much ^__^) and it'll give me something to do over the winter months.
The first thing I need for this project is a symmetrical power supply to give me positive and negative voltages, because practically everything synthesizer related needs a dual powersupply.
I used the LM317 and LM337 for this project because they are easy to work with and fully protected inside against short circuits and over heating. The LM3XX series regulators can deliver up to 1,5 Ampères.
This is the schematic I came up with and it works very well: (click picture for full screen view)


(STRIPBOARD LAYOUT FURTHER DOWN THE ARTICLE!)

I based this schematic on the datasheet schematic for the LM317 and LM337. It calls for a 5K trimpot but I used 10K's because that's what I had and it works fine. Afterall you can trim it down to below 5K just as easy, so no problem there. The values of the electrolytic caps I'm using is way over the top compared to the datasheet but it does help stabilize the voltage especially because some modules, like Sample and Holds or Sequencers can pull a pulsed current from the powersupply. Here's a screenshot of the original schematic from the Texas Instruments datasheet:



The transformer you use for this power supply must be about 2 to 3 Volt higher in output voltage than the needed output voltage of the powersupply. There is going to be some voltage drop over the voltage regulators and the diode rectifiers that must be accounted for. However the voltage will also be higher than the transformer delivers when it comes out the diodes so we must also account for that. I find 2 x 17Vac secondary outputs to be about the sweet spot. You can go a bit higher if you like.
In the schematic above you can see that the output capacitors are 1000µF. They don't have to be this big. 100µF will do nicely too. The caps discharge through the 220 Ohm and 10K potmeter when you switch off.
If you need a power supply that can handle more current, say 10 Ampères for instance, then you can easily adjust this circuit by adding a 2N3055 on the positive side and an MJ2955 on the negative side. You can even put more power transistors in parallel over eachother to get even bigger current specs. Move the capacitors that come after the LM3XX's past the transistors but put an extra 100nF capacitor between the base and the null to suppress transients etc. Use Google to find more specific schematics for that, if you need it.

But for a modular synthesizer those alterations are not needed because the individual modules I'm going to build won't draw much current at all. It's mostly in de 20 to 80 milli amps at most. The only transformer that I had that was big enough for this project didn't have a center tap but it had two independent secundary outputs. One at 21V and one at 17V. This was close enough. I connected one wire from each output together to make a center tap and made a circuit board to build the powersupply on. (Make sure when you connect two secundary windings together like this, that you measure the AC output before proceeding with the next steps. If you connect the wrong wires the 2 voltages will try to cancel eachother out. It won't damage the transformer but you won't get any voltage out.) I didn't have etching fluid anymore so I cut different islands on the copperside of the print using a dremmel tool with a milling bit. I had an old 25 Amp. Graetz diode bridge in my collection of components and after I drilled all the holes the build was pretty straight forward.
I used little trimmer potentiometers directly soldered to the print to set the voltage with. Once you set it, you don't have to touch it again but if you want to make it continuously adjustable you can of course use 10K panel mounted potentiometers with a knob. I put some LED's on the output side to indicate that all is working well. The difference in brightness will indicate if the voltages are set differently from eachother. You could also go the luxurious route and use two panel voltmeters to indicate the voltage but as I intend this to power my synthesizer that will not be necessary.
But if you build this as a stand-alone power supply, it would be a good idea to use two meters on the outputs.
The power supply as described above outputs between 1,5 and 25 Volts (dependent on the transformer you use) at a maximum current of 1,5 Ampere. If you want a more powerful version you could use the LM3XX regulators to drive a 2N3055 and a MJ2955 NPN and PNP power transistors as mentioned above and then you can draw up to 10 to 15 Amps. Be aware that the pinouts for the LM regulators differ from eachother. The correct pin numbers are noted on the schematic above.
The ripple is very low on the output. It's actually better than some you buy ready made. Make sure the electrolytic capacitors you use are rated 50 volts or higher. (I used 35V ones and that seems to work fine too but don't go any lower!) and make sure they are oriented the right way. Plus on + on the positive side and plus to ground and minus to negative voltage on the negative side. The voltage at the secundary directly after rectification can go up to 10 volt higher than the AC voltage from the transformer. Don't present more than 35V to the input pin of the regulators and be sure to use big heatsinks on both of them.
I've measured the ripple and noise of the power supply using the method outlined by Dave Jones from the EEVblog on YouTube in his video EEVblog #594 - How To Measure Power Supply Ripple & Noise
I used my simple home build 'brute force power load' described in the article from januari 2017 and under a load of 1 Ampere the Ripple Voltage Vrms was 6mV and Vpeak-to-peak was 10mV. Those are very good results.

Here's the stripboard layout for the powersupply. There are minor differences in values and components because this layout is based on an eBay kit, instead of the above datasheet schematic but it works perfectly, I assure you. You can mount the voltage regulators on a single big heatsink but they must be electrically insulated from the heatsink and eachother.

[NB: 11-Sept-2021 I just built one of these again today using this layout and it worked rightaway.]
You can leave the 10µF electrolytic capacitors over the trimpotmeters out. You don't need to include them. I put them in as an added protection for the LM3** voltage regulators but they are not needed. They are also not included in the schematic above.   


Here are some pictures of the first powersupply . As you can see there's some room left on the print for extra voltage regulators to get other voltages from the same supply:




I added the inductor coil in series with the 'Null' pole to suppress any high frequency noise. It's just something I added as a test but you can ignore that.

It's a week later and I finished the power supply as I need it for my synthesizer project. It now has -15/0/+15V,  -12/0/+12V and -5/0/+5Volt. I looks a bit of a mess as my projects usually do but it works just fine. Here's a picture of the finished psu:



Obviously you can't plug in all modules in the same holes so I've build a power bus system to which I can connect every module I build. It's a bit crude and I use a lot of hot-glue to stick it all in place but it works just fine and it will all be invisible once the case is ready.



Below you see the connector I build. The pins carry the following voltages: From top to bottom in the pic below, the top 2 pins are ground or 0V. Then I took out two pins and stuffed the holes in the female connector with hot glue. This is to get an a-symmetrical distribution so you can't put the connector in the wrong way around. Then there's -15, -12 and -5 Volt and then we get +5, +12 and +15 Volt. I kept the plus and minus pins as far away from eachother as possible for safety reasons.





PART TWO. SECOND POWERSUPPLY for stage two of my synthesizer.
So, as I'm writing this we're 6 months on in the synthesizer build and I'm about to add a second stage to go on top of the synthesizer I have already built.
So I need a second power supply. The first design as seen above works so well that I'm repeating it for the second stage with a few minor changes. I'm using multiturn potmeters for the LM317 and 337 voltage regulators so I can set them very accurately. In fact, the one I just built has plus and minus 15.00V that's accurate to 1/100th volt. I'm using all LM3xx regulators for this because I have a lot of them and because their voltage doesn't drop if you pull more current from them which is important because otherwise the VCO's would become out of tune. I'm using the same schematic as above and I made separate prints for every stage of the supply. Here's the stripboard layout:



Bridge rectifier print. (Don't forget to cut the copper strip underneath the fuses and to tin all the copper strips that carry current.). You can hang more than one voltage regulator print off of the bridge rectifying print as long as the transformer and rectifying diodes can handle the maximum current of the combined regulator prints.


Voltage regulator print. (Again, make sure to tin all the copper strips that carry current). 


(Last revised: 30-Jan.-2021: Corrected connection of negative voltage indicator LED.)


SOME NOTES ON THE DIFFERENT COMPONENTS TO USE:
Don't get confused by Capacitors being of a different value on the stripboard than on the schematic. The electrolytic caps on the Rectifier board are the big ones. They can be 1000µF to 2200µF or even higher and they do the main ripple suppresion. On the regulator board the electrolytic caps can be smaller, like 100µF because the main ripple suppresion has already been done and these are there to suppress noise and such. 100µF is enough for that.
Diodes also. The diodes around the voltage regulators are simply safety valves. Their purpose is to prevent the output having a higher voltage than the input, which could damage the regulator. The circuit will work fine even if you leave the diodes out. You can use any type of diode you wish 1N4148, 1N4001 upto 1N4007. It doesn't really matter. The diodes on the bridge rectifier however must be types that can handle at least 100V and 1,5 Ampères. You can not compromise on those but there are many different types you can choose from so I didn't specify which type number of diodes to use. You can also use a Graetz Bridge Rectifier, that's 4 big diodes in one case. I saw a 1000V/4Amp one for 50 cents on eBay. Those will work fine and they won't even get warm.
Like I mentioned before, the voltage regulators must be mounted to a heatsink. Either use two separate heatsinks (make sure they can't touch eachother) or use one big one for both regulators but in that case you MUST isolate the regulators electrically from the heatsink otherwise you'll get a very short lived microwave oven with integrated laser lights. ^____^
Naturally the voltage regulators do not have to be mounted on the print itself. You can mount them on the backside of a front panel, using that as a heatsink, or on the side of a metal case you're using and then you can wire them up to the print with normal electrical wire. Use your imagination but do make sure the regulators are not electrically in contact with eachother, otherwise, boom!
The LED's are simply there as a visual indicator that the circuit is under power and they are not critical to the working of the circuit, so you can do without them if you wish. Use 15K current limiting resistors with the LEDs though, because lower values can get hot.

Should you have problems like not getting the right voltages out of the powersupply then check your resistor values. Are you sure the 220 Ohm is not a 220K? This has happened on at least two occasions I know of that's why I'm writing it here as a reminder ^___^

Here are some pictures of the finished power supply. Tinning all the copper strips that carry current is important because they get very thin around the holes in the stripboard. 
I've mounted the whole powersupply on a long piece of MDF ready to accommodate the power-bus system I need to build.






I found some very old vintage diodes with a metal case which I think look very cool and they work fine. They are sturdy too because I had some short circuits in testing and the fuse went 2 times but the diodes didn't mind and I use slow fuses too so they did get some current through them.
Use plenty of heat conducting compound on the LM's. The electrolithic caps are 1000µ each; all four of them and that's all the capacitance I put in. 1000µF in the rectifier and 1000µF on the output side of the voltage regulators. The rectifier caps have 10K 1Watt resistors over them to make sure the are drained of voltage when the powersupply is switched off. (It's 2K2 in the picture but they were getting slightly warm so I changed them for 10K's)


SIMPLE FIXED VOLTAGE POWERSUPPLY  using the 7812 and 7912 voltage regulators.

Finally I want to close off this article with a very simple powersupply that uses fixed voltage regulators. The 7812 for positive voltage and the 7912 for negative voltage. These can do up to 1,5 Ampères but I wouldn't use it for more than 1 Amp. otherwise they get very hot even with heatsink.
Btw, you must use these on a heatsink too, just like the previous designs, and you must make sure they don't electrically touch the heatsink if you have both regulators on one heatsink. Otherwise you get a short circuit and a lot of magic smoke. The circuit schematics for this one can be found by clicking here

Here is the layout I made for this PSU. It's very small and can fit anywhere. The LEDs are there to indicate if power is present on the outputs. You can mount those in a panel if you make a panel for your powersupply. I always mount leds like this near the ON/OFF switch for the PSU.


(Last revised: 24-Aug-2022: Corrected a mistake where the LEDs were not connected to ground properly, thanks to a observant reader. )

Okay that's it for this one. If you have any questions you can leave them in the comments or post them on our special Facebook Group for this website, where we have a cool little community who will love to help you out.

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!

53 comments:

  1. Hi Eddy. I'm struggling to find a center tap or a double winding transformer. Would two transformers wired in series work too? Thanks

    ReplyDelete
    Replies
    1. Yes that would work too. It would be best if they were two the same transformers but it's not necessary. Make sure when you connect them that you measure the voltages. If you connect them wrong the two will cancel eachother out. In my first powersupply I use a transformer that had 21V and 17V windings and connected those together. Worked fine.

      Delete
  2. Ok, thank you very much. They're both 24v in fact.

    ReplyDelete
    Replies
    1. Okay, measure the voltage after it has been rectified by the diodes and make sure it's not too high for the voltage regulators to handle. The diode rectification can add upto 10V to the voltage output directly after the diodes.

      Delete
  3. Hi Eddy, I need your help once again. The power supply seems to be working fine, and I can regulate it using the trimmers to get ±15V. However, when I connect a load to it, the voltage goes down significantly. It happens both with the positive and negative output. Do you have any idea why this is happening? Thanks in advance.

    ReplyDelete
    Replies
    1. It looks like your transformer can't deliver enough current. I can think of no other reason why that would happen.

      Delete
    2. I think the problem is in the board with the regulators, because the ac voltage of the transformers and the dc voltage after the rectifier stage don't go down. Also, I've tried powering the regulators with a 11.5V power brick I had lying around, to take the transformers out of the equation, and when I connect a 220 resistor to act as a load (which shouldn't be taking too much current) the voltage still drops from around 8.5 to 5. However, the funny thing is that this happens to both the positive and negative outputs, so it's not probably due to a faulty regulator. I've gone as far as to desolder the diodes, but still no improvement. I have not the slightest idea why this is happening, but I'll keep on trying to figure it out.

      Delete
    3. That's a weird problem. I took the design of my powersupply mostly from the datasheet of the regulators and I built 3 of them sofar and they all work fine. I'd be interested to know what the cause is. I hope you figure it out and please keep me informed.

      Delete
    4. Well, I finally found the problem and I have to say that I'm profoundly embarrassed. I mistook a 10k resistor for a 220 one! In my defense they only differ in one line, red on the 220 one and brown on the 10k one. Nevertheless it's finally working as intended now, so please forgive me for bothering you with such a stupid mistake :')

      Delete
    5. Ah that explains it. Don't worry about it. Can happenen to us all :) I'm glad it works now.

      Delete
  4. Hi Eddy, I'm having an issue setting the voltages for this unit. I was able to set them at one point, but once I connected a load to it there was a significant drop. I read the other comments on this page and saw what one possible solution was, initially I had used a 68ohm and a 150ohm resistor in series to make the 220 ohm resistor since I didn't have the exact value, I have since acquired some 220 ohm resistors so I tried making the replacement. After making the replacement, I have found it to be impossible to set the negative voltages at the desired level. I can bring the voltage up to zero and values near zero, but as soon as I try to bring it down to the desired -12 or -15, it immediately jumps to something like -24 volts. During one attempt the lm337 actually completely failed and smoke came out and the fuse also blew.

    As far as I can tell my board is exactly like the one in the provided layout. The transformer I'm using is a 35vct rated at 1.5A, and I measure about 20v AC coming out of each lead if that changes anything. Any ideas? Thanks.

    ReplyDelete
    Replies
    1. I should add: after rectification, I'm measuring +-27 volts DC.

      Delete
    2. That is weird. You must have done something wrong, it almost looks like one of the protection diodes was mounted in the wrong direction. I can only advise to take a printout of the schematic and check the positioning of every component and see if there's something wrong.

      Delete
    3. Ok so I took a look at the diodes and triple checked the layout to the best of my abilities and didn't find anything out of place. I tried testing each negative voltage circuit independently by only suppling it from the rectifier one at a time and the result was the same. Would this lead you believe the issue could be with the rectifier or supply from the transformer?

      Delete
    4. If the transformer can deliver 1,5 Amp it can't be the problem. Are you sure the potmeters are okay that set the voltage?

      Delete
    5. Ok. I did try swapping out the pots with new ones, no change. What confuses me the most is that it happens on both the -12v and -15v circuits, but not the positive ones. Heres a couple pictures just in case Im missing something blatantly obvious: https://imgur.com/a/jigMaq2

      Delete
    6. Well I had a good look at your pictures but it all seem good to me. I double checked my layout but that is okay too. Try taking out the diodes around the negative voltage regulator and leave them out. They're not essential for the working of the circuit. See if that solves anything.

      Delete
    7. Thanks. Just tried it but it doesn't seem like anything changed. Just like before the fuse blew moments after I plugged it in.

      Delete
    8. OK, so it seems like the fuse blowing is related to how the pots were set when I plugged it in, I had them set all the way up, but when I start with them open, the fuse doesn't blow and I am able to start bringing the voltage towards the value I want before it suddenly jumps to zero. Considering the input voltage is higher than the specified 17v, do you think a higher value pot could solve this issue?

      Delete
    9. I went ahead and tried a 20k in place of the 10k, no luck.

      Delete
    10. Well, I'm using the datasheet schematic for the regulators and the potmeter needs to be that specific value. This is mystery here. Have you tried new regulators? Did you go over the schematic and checked it against your component placement? (I bet you have). I can't suggest any other option.

      Delete
    11. I am out of ideas as well. I think I'm just going to start over with all new components and hope for the best. Thanks for your help.

      Delete
    12. You're welcome. Sorry we couldn't make it work.

      Delete
  5. Hello. I have same transformer like you but don't know how to connect these 7 wires on primary side for 220v, can you provide some picture or circuit please. Thanks.

    ReplyDelete
    Replies
    1. I'm sorry but I can not photograph the transformer because it is built into my synthesizer now. But what you can do is measure the resistance between each of the wires on the primary side and choose the ones with the highest resistance. That will be the 240V connection.

      Delete
    2. The resistance will be about 39 Ohms.

      Delete
  6. Hi Eddy, will this power supply work on 120v?

    ReplyDelete
    Replies
    1. Yes of course as long as you have a transformer that has a 120V primary winding then there's no problem.

      Delete
  7. Hi Eddy, I’m planning on building many of the modules on your site and have been building up a decent stock of parts but would rather buy a power supply. Is there one that you can recommend to me?

    ReplyDelete
    Replies
    1. Hi Sammy, that's alright, I deleted the other two identical remarks. As for a powersupply I can recommend. There are some good kits available on eBay. I recently got one of these for a project and they work very well. It's got dual voltage and you can set the voltage with a trimmer. https://www.benl.ebay.be/itm/LM317T-LM337T-317-337-Power-Supply-Dual-Voltage-Regulator-Adjust-Board-Kits-L-/232981506790?hash=item363ec7aae6

      You can find similar kits all over the internet. You do need a good transformer though with minimal 2 x 15 Volt with a center tap but you can get those from old radios or stereos at second hand shops. I never buy new transformers.

      Delete
  8. I don’t mind buying a transformer due to the convenience of it coming to my door step would this do?
    Jameco Valuepro 112512.-R Power Transformer, 24 VAC 2 Amp 117 VAC Wire Leads https://www.amazon.com/dp/B00B8865AC/ref=cm_sw_r_cp_api_glt_fabc_DFR4Q4MSPB4SS5M2H861

    ReplyDelete
    Replies
    1. No you need one with three secundary wires, like: 2 x 15V AC

      Delete
  9. I realized that I have a meanwell rt-65b which is +/-12v and also has a 5v output so I’ll just use that since I already have it. Was originally intended to put in a eurorack case but I can just order another when I get the motivation to pull all the modules and modify the case. Problem solved!

    ReplyDelete
  10. Have you thought about doing a wasp filter layout? I’d be super hyped on that! Stoked on building what you have available too but I’d be very excited if there was a wasp filter build

    ReplyDelete
    Replies
    1. I think I will do w WASP filter in the near future. I'm curious too. :)

      Delete
  11. Sweet! I have the doepfer Wasp but would like to build it and compare them

    ReplyDelete
    Replies
    1. There are stripboard layouts for the WASP filter online. I won't start building again at least until autumn. I'm having a little summer stop.

      Delete
  12. Hello Eddy,

    I've made a start and nearly finished this power supply, using bits I'd already got. Worked first time I'm pleased to say. https://photosynthvideo.blogspot.com/p/power.html It'll be in a separate box and connected to the case with a 3 pin 12mm Aviation plug and socket.

    Cheers and many thanks - Jim

    ReplyDelete
    Replies
    1. Hi Jim. That looks great! I'm glad you got it working without any problems. A separate powersupply is very handy. That way you can also use it to test new projects. I built myself a separate dual powersupply too and included a Thomas Henry VCO (the first one I built) as a function generator and an LFO for frequency sweeps. Now I have a complete desk top function generator slash powersupply for testing. :)

      Delete
  13. Hi Eddy, I already build some modules from your designs and they work really fine. I just built the power supply. I noticed that there are caps of 10uF over the two pots. I only see them in your first layout. What is their function? The supply seems to be working fine without them.

    ReplyDelete
    Replies
    1. Hi Ton. You can indeed leave those 10µF caps out. They are used as added protection for the voltage regulators (see fig. 17 of the LM317 datasheet pdf) but they are totally not needed for a synthesizer powersupply and they are not in the schematic either. I guess I put them in to test it, I can't remember anymore. I'll make a note in the article that they can be left out. Thanks for drawing my attention to it! I'm glad everything works so well.

      Delete
    2. Thanks! And thanks also to put all this info out in the open. I like your designs. They are changeable. Every time I order parts they have different dimensions as the ones I see in your pictures so its nice that there is room for change :)

      Delete
    3. Thank you so much! Yes I set most designs up with enough space for changes. That makes troubleshooting easier too, when necessary. :)

      Delete
    4. Unfortunately I have the same problem as those I read in the thread...I rimmed the pots to a perfect + and -15v. When I hooked up a vco I first thought that the vco was broken. So when a load is connected the power drops a lot . I guess one vco doesn't pull much power. I have no clue. Everything is connected in the right way. I checked it many times.

      Delete
    5. Are you sure the transformer can deliver the right power?

      Delete
    6. Yes, I also ran all my modules on another powersupply from a diy mixer I built years ago. It has a much smaller transformer and everything ran really well. I built more supplies and I never had a problem like this.. Could the problem be due to the 2 trimmer circuits? The circuit is very straight forward so I have no clue at this moment.

      Delete
    7. This is a difficult one. They must be a mistake made somewhere because these layouts have been used by som many people now. It's simply the datasheet circuit I'm using. Are you sure your transformer has enough over voltage? For instance to get 12V out you can't get away with putting 12V into the voltage regulators. The input voltage needs to be about 3 volts higher than what you want the output to be.

      Delete
    8. It's 18V... Checked everything, measured everything. I will check the internet for more info. If I find something that solves it, I will let you know :)

      Delete
    9. Okay thanks. This is a weird problem. I hope you can fix it.

      Delete
  14. I don't know how much current each of my modules draw, but i'm over 30 modules deep right now, I guess it would be smart to build a second supply or one of those transistor current boosters right?

    ReplyDelete
    Replies
    1. Over 30? Yes it would be a good idea to build a second powersupply. I have about 15 modules hanging from each of my powersupplies. Every row has its own supply and I have a temperature meter connected to the heatsink of one of the LM317's and that goes to 65 degrees Celcius quite fast. Then it switches on a fan and then the temperature goes down again and the fan switches off again. So it goes on and off. That's not necessary though. Just something I built in ^____^

      Delete
    2. Yeah I dont know why I didn't consider this earlier on, It has been going strong for quite a while now at about 30 modules but i can imagine it overheating to its death soon so I'll unplug some things when I play it until I have a second supply

      Delete

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