[splaz]

Okay so I want a bench supply however commercial ones are too expensive and used ones are hard to come by and too expensive to ship due to weight.

So thought I'd build one.

If anyones found some good designs in their internet travels let me know.

I've found this one so far which seems fairly impressive in terms of voltage and current output.

http://www.uoguelph.ca/~antoon/circ/ps3010/ps3010a.html

Want dual polarity so I'd build 2 into the one enclosure. Also searching here on the subject tangent said 5V is useful to have in power supplies. I have a spare 338K lying around, would it be worth building a circuit and setting that to 5V and adding it in ?

 

[sbyers77]

I am not sure how much current you need, but you can definitely make a bench top power supply out of an old ATX computer power supply. The supply's are +3.3V, +5V, +12V, and (I think) -5V, and -12V regulated DC. They are cheap to come by and there are plenty of guides online that can show you the conversion. Example 1, Example 2.

 

[SiBurning]

Simple power supplies are easy to diy. In amplifiers, you probably end up building these all the time because the supply is such an integral part of the amp. I just keep parts around and build them as needed on a separate breadboard. It's useful to make a few cheap low voltage ones just to have around. 5V is definitely useful, especially if you use any digital circuits. So is 3.3V.

In general, though, I'm not a big fan of diy with supplies. The cost/benefit isn't worth it. For many amps, you end up having to design or build a dedicated supply to go with it anyway, so it's not a huge thing.

You should consider buying a used supply. A good bench supply includes things that would cost a lot more to DIY. The main things are solid build quality and current limiting. There's always a bunch of HP power supplies for sale on eBay for $50-100us, plus maybe $20-30 shipping. You could spend a lot more than that on the case, switches, connectors, displays, and transformer alone. It's like getting the electronics for free. If you really need something special or precise that you can't buy used for cheap, it might still be worth it to buy a used piece, gut it, and add your own electronics.

The reason I recommend HP is they have so many options. I bought an HP 6228B dual supply for $130us including shipping. Does dual 50V with 1A current limiting. Be careful with the cheap DIY designs here--many of them drop the voltage as current limiting kicks in. Eventually, I'll probably pick up a couple of other supplies with different specs, maybe one with higher current, or just a couple of cheap $25 Tektronix T500 plug in modules. The thing is, they're not all that useful. If I can send 38V to the main supply, I'll have to build the low voltage supply for the early amp stages anyway, so why bother with a second unit?

 

[Vertigo Acid]

Quote:

Originally Posted by sbyers77 /img/forum/go_quote.gif I am not sure how much current you need, but you can definitely make a bench top power supply out of an old ATX computer power supply. The supply's are +3.3V, +5V, +12V, and (I think) -5V, and -12V regulated DC.

ATX spec recently (since 2.03 ?) has dropped -12v and -5v. If you're going to retrofit an old one, make sure it has a blue and a white wire, for -12v and -5v, respectively.

 

[splaz]

Quote:

Originally Posted by SiBurning /img/forum/go_quote.gif Simple power supplies are easy to diy. In amplifiers, you probably end up building these all the time because the supply is such an integral part of the amp. I just keep parts around and build them as needed on a separate breadboard. It's useful to make a few cheap low voltage ones just to have around. 5V is definitely useful, especially if you use any digital circuits. So is 3.3V. In general, though, I'm not a big fan of diy with supplies. The cost/benefit isn't worth it. For many amps, you end up having to design or build a dedicated supply to go with it anyway, so it's not a huge thing. You should consider buying a used supply. A good bench supply includes things that would cost a lot more to DIY. The main things are solid build quality and current limiting. There's always a bunch of HP power supplies for sale on eBay for $50-100us, plus maybe $20-30 shipping. You could spend a lot more than that on the case, switches, connectors, displays, and transformer alone. It's like getting the electronics for free. If you really need something special or precise that you can't buy used for cheap, it might still be worth it to buy a used piece, gut it, and add your own electronics. The reason I recommend HP is they have so many options. I bought an HP 6228B dual supply for $130us including shipping. Does dual 50V with 1A current limiting. Be careful with the cheap DIY designs here--many of them drop the voltage as current limiting kicks in. Eventually, I'll probably pick up a couple of other supplies with different specs, maybe one with higher current, or just a couple of cheap $25 Tektronix T500 plug in modules. The thing is, they're not all that useful. If I can send 38V to the main supply, I'll have to build the low voltage supply for the early amp stages anyway, so why bother with a second unit?

Well I did say I cannot easily find commercial units over here used, no luck so far atleast, found some cheapish new ones but those are of questionable quality. I would get them from elsewhere but because they often have large heatsinks, trafos, possibly metal cases etc. they are quite heavy and so well, shipping is extremely costly to the point where it isn't really worth getting them from other states/overseas. I'm having the same trouble at the moment trying to get an old CRO, now that I definitely can't DIY up. I can get them from the eastern states but once you add in shipping it isn't really worth it. Many places want above the items cost in shipping for such heavy equipment.

This DIY design does have current limiting and is 30V and up to 10A.

I want an adjustable one, however I have used old ATX power supplies in the past where I just needed 12V, but I honestly went through a whole pile of them before I found one that wasn't dead, and even then the one that worked had substantial dips in voltage when loaded, so I've had problems with reliability on the cheap ones and they're fairly noisy as well, plus as said, not variable although you can get the common voltages from them.

 

[Garbz]

Just read the datasheet for the LM317 regulator. Under application notes there are examples of bench supplies with adjustable current limiting and voltage.

 

[amb]

See my post in this headwize thread about the problem of DIYing a linear bench supply. It touches upon some issues that you may not have thought about.

 

[SiBurning]

Most computer supplies are terribly unstable. I decided to only use PC Poewr & Cooling supplies from now on because I've been fed up with lousy supplies, and would rather pay the $200 and be done with it. The other problem with computer supplies is they're switching, which means tons of high frequency noise.

I didn't notice before you're from Australia. If you don't mind a really odd idea--and you should think about it again after you stop laughing--try finding a veterans hall or radio club and asking if any of the old timers have any radio equipment in their garage or knows anyone who has gear they don't need anymore. I bet with a little patience you end up with a garage full of surplus that's serviceable at no cost. You might even get lucky and find an active ham that has decent stuff that's not 40 years old. Even if the stuff isn't the best, you can gut it and grab some really awesome transformers, cases, and other parts. But I bet a lot of the 50 year old stuff is still quite serviceable for the basics. If you want to be more selective, no problem taking this route: you just need a little more patience. Eventually, someone will know where to look.

 

[splaz]

Thanks for pointing that out amb. I was considering a tunnel heatsink with push-pull fans however that was just to keep it nice and cool, in hindsight as I was forgetting it was dropping that much voltage accross it, that's probably a necessity in the first place.

Well I'll continue my search for a used one, that suggestion did make me laugh a bit SiBurning. Actually a really good idea though, getting an old timer to pass some gear on to a young whipper snapper such as myself.

I actually was looking at getting an amateur license for HF radios so I may be able to get that and score free stuff all in the one go.

 

[Garbz]

SiBurning while computer supplies are less than stable I find them very exceptable for their purposes as a prototyping supply. Follow them with a larger capacitor and some very minor filtering, I have used this for all of my projects thus far. No it is not ideal but it is a great way to see if a circuit is working before building a proper psu. The downside is ofcourse no adjustable current limit.

 

[SiBurning]

Quote:

Originally Posted by Garbz /img/forum/go_quote.gif SiBurning while computer supplies are less than stable I find them very exceptable for their purposes as a prototyping supply. Follow them with a larger capacitor and some very minor filtering, I have used this for all of my projects thus far. No it is not ideal but it is a great way to see if a circuit is working before building a proper psu. The downside is ofcourse no adjustable current limit.

Excellent point. I already said how it's useful to have some cheap supplies--this should've been obvious. How do they sound? Or is it just for the quick & dirty stuff?

Anyway, for current limiting you might be able to use a light dimmer. Or is that getting too cheesy?

(Time to toss the banana into the monkey cage.) Working with voltage lag is probably a useful skill. The schematic for my Kenwood says the supply is 50VDC, (which is actually impossible given 33vac off the transformer) but at some points it's really as low as 41.3V, and after doing the math, I'm pretty sure that's how it was intended to come off the shop floor. I talked to a repair guy from those days who says that was typical for equipment in the 70s. It's kind of hard to know if the things working okay when the schematic says a point should be at -1.7 and you read -0.6.

 

[vixr]

I bought one of these from a guy on ebay... It is a kit and works really well
http://cgi.ebay.com/Bench-Top-Power-...QQcmdZViewItem

 

[tangent]

Don't discount a good current limiting feature. I can't tell you how many times this has saved one of my circuits. Something goes wrong, current races for the ceiling, but is stopped by the bench supply. You immediately see this on the supply's front panel, and can go looking for what caused it. Without a current limiter, the circuit would have cooked itself, and you'd have to spend a lot of time rebuilding.

It's tolerable if the output voltage falls in order to meet both the output current and voltage requirements. A lot of the lower end units just switch from constant voltage to constant current mode when the current limit is hit, allowing voltage to sag. Although not ideal, it does still serve the main purpose, which is to protect the downstream circuit. It might cause secondary problems, though, like a big DC offset in a headphone amp circuit, or processor resets in a microcontroller circuit.

It's better if the supply is smart enough to maintain the requested voltage while it keeps the current limited, but this is less common than you'd think.

 

[Garbz]

How's it sound? Well how you would expect. It's average, there was nothing terribly wrong with it asside from a tiny bit of noise generated on the 12v line with a fan. The switching noise was outside the audible range. But when powering my DAC for a week or so while building the PSU it still sounded better than my computer soundcard.

But the fact of the matter is for $18AU ($13US) you can get a new 400watt ATX psu. It is no bench supply, but then it doesn't cost that much either.

 

[mono]

I'd beware of such cheap ($13/400W) PSU, they often are of questionable ratings, build quality, and might lack some AC line isolation & filtration.

If the supply uses a fairly common PWM controller like TL494, the feedback can be reverse engineered enough to bend it to your will - a little. Things like changing the feedback to entirely focus on the 12V rail so you have highly regulated power from that w/o dips or any load necessary on 5V rail, even more accurate if you use a remote sensing lead at the connector. That "12V" can also be varied some, if you used a pot on the feedback you can get a range around 5-16V, thojgh if the other PSU rails are monitored for over/undervoltage in addition to current you'd need to disconnect them from the comparitor as well. Download the TL494 spec sheet and it will give you some ideas, basically the idea is just removing the 5V feedback path resistor that's parallel to a 12V, and increasing the 12V resistor so it's the primary.

It'll still be a switching supply though, noise often centering around roughly 50KHz.