Jump to content

+12/-12/+5 bipolar PSU circuit


fussylizard
 Share

Recommended Posts

I'm trying to design a PSU board for my MB-SEQ w/ and AOUT-NG and SSM filter. I swiped the basic schematic from the MB-808 and thanks to Altitude and nILS made some tweaks and simplifications, and I also tried to make it easy to build (e.g. just using 2200uF filter caps for everything instead of more exact cap values- one type of electrolytic cap means less confusion, etc.).

If it is not clear from the schematic, the jack at the left is for a center-tapped transformer, and there are multiple SIL headers for output power.

I've built it on a protoboard and it seems to work, but I had a few questions I was hoping someone might be able to answer:

1) Do I really need the protection diodes D1/D2/D3/D4? As far as I can tell this comes from an old app note from the 1970's to protect the regulators, but modern 78xx/79xx regulators are pretty indestructible and I don't see these in many other designs, so can those be removed?

2) Can I tie AGND and DGND together? It seems like I should, but I didn't see that in the MB-808 PSU so I thought I should double-check.

3) What's the purpose of D5? It was on the MB-808 but I wasn't sure of its purpose. I'm guessing it helps prevent some sort of current flow between the 5V and +/-12V sections?

Thanks for any insights!

Oh and BTW I know I can probably buy something off-the-shelf that would work for me, but I wanted to make my own circuit board in Eagle and get it made (at BatchPCB). It's been fun so far. Once the design is proven, I can post the .sch and .brd files. I could probably be talked into a mini-bulk order for PCBs if enough people are interested.

Regards,

C

post-6425-0-37514700-1311393125_thumb.pn

Link to comment
Share on other sites

that's a great job to have produced a so clean schem, especially for beginners

I've built it on a protoboard

me too, a long time ago for general purpose then integrated to my MB CV (+12/-12V analog, 5V digital). So i need to do another one (based on variable VR)

It works perfectly. it has the quite same schematic BUT :

- i didn't used D1 D2 D3 D4 as there are LED that do the same : abort negative pulse in case or short circuits or similar, showing that a voltage is present (and doing a load by the same occasion in order to mesure the exact voltage)

- I used only 2 caps of 1000u (+ one 2200 on the core) before the VRs, the rest being 10u, 330n and 100n. Long observation with an oscilloscope showed perfect linear voltages.

Connecting DGnd and AGnd will create ground loop in the final circuit with Core and AOUT_NG

Making it on veroboard goes very fast, the longest is to drill additional fixing holes for the transformer ; moreover this permits to adjust the size of the 2200u capa as you know that depending of the voltage, legs space is different. The necessity to adjust the voltage of capa is important as the cheapest is to recycle an old transformer from a VTR for example.

I restored last days a Juno 106. You should copy its powersupply and then do a batch : there are 5V in analog and digital, +15/-15 in analog, and a 9V line (what for ??). there are trimmer to adjust voltages. It has fuses (5 in total!) that saved my life and my VCF/VCA circuits while i was experimenting into. There is also a retrocontrol circuit into (as on Switch PSU) but with a transformer ! So cool :)

in french : http://www.sonelec-musique.com/electronique_bases_alimentations.html

post-4619-0-91335900-1311399000_thumb.pn

Edited by julienvoirin
Link to comment
Share on other sites

The protection diodes are a good idea, especially if you are using unusually large capacitors after the regulators, as you are in this case. The "dangerous condition" is after power is removed, the voltage at the output of the regulator will be greater than the voltage at the input.

Why did you do this on some regulators but not all of them?

AGND and DGND should be connected in one place. and the power supply is usually that place.

D5 drops the voltage to the 7805 just a bit. this means it will run just a bit cooler, and the heat from D5 is exactly that amount. In the end, net heat is the same either way.

A suggestion: Instead of using 2 pin SIL for the +5 outputs, use 3 pin with the center grounded and BOTH outside pins at +5 volts. This means less chance of plugging it in backwards, but more likely to get it confused with the 12 volt connectors..

Have fun!

LyleHaze

Link to comment
Share on other sites

I agree with Lyle. It's a good idea to do that. If you for instance look at the schematic for the Roland Juno 60, you will see that they do the same. And i would do all the regulators.

Just a warning you will need a serious heatsink on the 7805. The transformer you will be using is at least 12VAC. Rectified it's close to 17V. As said there will be some losses, but it will still have to eat away something like 10V x the current. So if your setup draws 1A in current, that will be 10 Watts heat dispensation.

Link to comment
Share on other sites

Is D5 really doing that much though? You could perhaps add a power resistor to knock the voltage down more before it goes into the 7805. Like the regulator, the resistor will throw off excess as heat but I found some ceramic ones at Radio Shack that have a large surface area and seem to be heat tolerant. Plus if you have a metal chassis you can just mount the resistor to it making the chassis something like a heatsink.

Link to comment
Share on other sites

I can't see any practical point for adding D5, either.

I agree that dropping something like 10V on the 5V reg is not too good an idea; I just tend to use 2 transformers if the 5V rail is going to be drawing a lot of current. A slightly more involved solution is to pre-regulate with a small switch mode circuit, it doesn't have to be that accurate, and will generate far less heat.

One thing no-one seems to do for this type of gear, is fit some kind of 'power fault shut-down'. I 'll try and do some details when I have time, but on my MB6582 power supply I will be fitting a 'crowbar' to the important rails, (this is a circuit that shorts out the power supply with a big thyristor if it goes over voltage); those SIDs are way tp precious to risk a PSU accident killing them. An having ground open an olld C64 supply, there is no way i'd trust one, long term, especially not with 2 disparate rails 'stacked'.

+1 on the protection diodes.

Link to comment
Share on other sites

If you insist on a linear psu, a power resistor is a good idea. You will need something like an 8 Ohm Resistor capable of handling at least 8 Watts. This way it will be possible to get the theoretical max of your 7805. If not the thermal protection will limit the amount of current flowing through the 7805.

A thyristor is overkill in this situation imho. If dealing with SIDS i would say yes, since they are pricey and vulnerable.

Edited by Shuriken
Link to comment
Share on other sites

That's exactly what I've done, only I used 2 4Ohm resistors in series with my old design. I don't have all the circuit protection stuff put together yet but you can look at my latest attempt here. That said, the linear PSU you can purchase mentioned in thread is compelling so I'm not sure which option I am going to go with at this point.

Edited by m00dawg
Link to comment
Share on other sites

Holy damn! What are you powering with -/+ 15V? I thought the AOUT stuff all ran on 12? Regardless that looks pretty pro, but I would expect nothing less from the likes of nILS :)

A modular PSU of course.... All modular stuff is +/-15VDC. But you can adjust the LM317/337 to the right voltage :thumbsup:

Edited by Shuriken
Link to comment
Share on other sites

that's a great job to have produced a so clean schem, especially for beginners

Guys, thanks for the info!

@julien - Glad you liked my first .sch layout!

Good idea looking at the Juno-106 PSU! I could find lots of examples on the web that were just bipolor PSUs, but I didn't find anything that also had a separate +5V off the same transformer.

Interesting that they have two +5V power supplies - one for digital (top section with IC1), one for analog (bottom section). The analog grounds for the +15V, -15V, and +5V are connected together per lylehaze's comment I think (see vertical line just to left of the VR2 bold label). Presumably they are separate for noise reduction.

The digital +5V, however, seems to come in on some separate winding on the transformer so I'm not sure what's going on there. There's also some extra circuitry above IC1 for reset that probably just triggers a reset pin on a microcontroller once the PSU is stable after being turned on.

Around IC2 there are some extra transistors and resistors that I don't know what they are for...will have to lookup the data sheet for that part if possible.

Doing a PSU board for others would be great, but I'll need some design help (<trek>I'm a coder not a circuit-designer, Jim!</trek>). And in the meantime, I want to finish my MB-SEQ. :-)

Link to comment
Share on other sites

The protection diodes are a good idea, especially if you are using unusually large capacitors after the regulators, as you are in this case. The "dangerous condition" is after power is removed, the voltage at the output of the regulator will be greater than the voltage at the input.

Why did you do this on some regulators but not all of them?

Because the MB-808 PSU doesn't do it on the +5V rail for whatever reason. :-) I agree it probably makes sense to add the diodes to the 7805 as well.

AGND and DGND should be connected in one place. and the power supply is usually that place.

Do you have any idea why the MB-808 doesn't connect the two grounds? See the attached image. It seems weird, but that's what is in the .sch. (All the extra caps at the right are just despiking caps on each chip AFAIK...they just put them in the PSU schematic.)

D5 drops the voltage to the 7805 just a bit. this means it will run just a bit cooler, and the heat from D5 is exactly that amount. In the end, net heat is the same either way.

Ah, right, duh!

A suggestion: Instead of using 2 pin SIL for the +5 outputs, use 3 pin with the center grounded and BOTH outside pins at +5 volts. This means less chance of plugging it in backwards, but more likely to get it confused with the 12 volt connectors..

Interesting. Not sure which way is better to go, but good to think about.

post-6425-0-53942500-1311454560_thumb.pn

Link to comment
Share on other sites

Just going through the other comments:

@shuriken, theprof, m00dawg, etc.

Regarding heat sinks, I agree completely. The +5V reg will definitely need them, plus maybe some diodes or power resistors (MB-808 only has the one diode).

Altitude suggested to me a while back using a drop-in switched mode replacement for the 7805: http://www.mouser.com/ProductDetail/Murata-Power-Solutions/78SR-5-2-C/?qs=sGAEpiMZZMukgiigmf73gFbB24B2O0wZ

Expensive, but no heat issues. Not sure about the noise (but it switches at 200KHz). Currently backordered, but I might give one a try.

@moodawg - nice board. You have a pic of it built? Why only space for the heat sink on the 7812 but not others? I attached my latest (non-final) board for reference. (I need to add the extra diodes.)

@nILS - I think you *way* undersized the heat sinks there! :-)

Link to comment
Share on other sites

Oh and a big thanks is in order to Altitude for all his help so far on my SEQ (I am using the same case as his...pictures to follow when I'm done!) and his .sch/.brd feedback along with nILS. :-) Thanks guys!

And thanks for everyone's comments so far! It would be nice to have an 'official' PSU for MIDIbox, but given that I'm a cut-and-paste circuit designer I'm not sure I'm the right person for that...

Regards,

C

Link to comment
Share on other sites

@moodawg - nice board. You have a pic of it built? Why only space for the heat sink on the 7812 but not others? I attached my latest (non-final) board for reference. (I need to add the extra diodes.)

My plan there was to fix the other regulators to the side of the chassis itself. You cannot do that to the 7912 without insulating it so I figured a heatsink would be better than since I didn't expect it to throw off too much heat. I haven't finalized the design yet as I am working on the control surface bits of the synth and I want to finish all that first. Despite that fact that they are massive, I do like nILS' design :) EagleCAD even had those type of heatsinks in its built-in library if I can find the part numbers of them to buy off Mouser.

As far as your attachment, I didn't see it in your last reply? I saw your previous one though.

Link to comment
Share on other sites

Not sure I follow on how to share them across rails. Can you elaborate?

Using 2200uF caps before/after the vregs on each rail is clearly overkill. I did it to keep the build simple. My earlier design had various 1000uF 35V caps, 2200uF/16V caps, 2200uF 25V caps, etc. I thought if others used the design it might be confusing, so I traded a little board space to use the largest caps needed on everything for consistency.

I've also thought about tweaking the layout to keep all the diodes in the same orientation to avoid those mistakes as well...

Link to comment
Share on other sites

The large electrolytic caps are generally to smooth out the AC voltage after it has been rectified. So you can likely just use the larger caps after the rectifier and share it amongst all the regulators. See my above deign for a bit what I'm talking about that.

To put it another way, you likely will want to use smaller caps around each of your regulators (check their spec sheets for the sizes) but probably don't need to have a big electrolytic cap before each regulator.

If you wanted to go above and beyond, you could use larger electrolytics and more of them but share them amongst all the regulators. I'm having trouble describing it without a picture I guess :) Some designs I have seen used 4x 1000uF capacitors to help smooth out the AC current amongst them. Supposedly that was better than using one big 4000uF cap though I've never put that to the test (I don't have an oscope to get a good reading for that).

Err I hope that helps :)

Link to comment
Share on other sites

Oh, I see on the series thing. I think the reason it is better is that the smaller caps can react faster, so you just use multiple to get the total capacitance you need.

@shuriken - 100uF is enough? I really should put a scope on it.

@all - At this point I think I'm going to go with my current version just to move on. Will let ya'll know how it turns out...Thx for all the input!

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
 Share

×
×
  • Create New...