Bench power supply

[6PTsocket]

I apologize if thas has been already beat to death. I have an assortment of power supplies, most fixed output. I want a little single output: CV, CC around 30 volts 5A give or take. No super critical applications; charging, led testing, seing if stuff will power up, etc. A switching supply is probably clean and fast enough. I see tons of stuff in the $55-90 range. Where I need advice is on the fine points. Some are only 3 digit, some are 4, one switched the meter from A to ma. Some have course and fine tuning some with only one knob. Is it multi turn? What I am concerned with is easy adjustment and enough meter to go from single digit ma. to A without having to use my dmm. I have seen reviews complaining about twitchy adjustment and and slow response. I also saw a Mastech liner 3A supply that they cut the price on to $89.  Any good?  I don't know if there is much difference in supply protection or quality of electronics or mechanical construction in this price range. The weight and size of a linear is not a problem. Any advice appreciated. Thanks.


Sent from my SM-G900V using Tapatalk

[cdev]

I like it when you can easily adjust the constant current limit value to a fairly precise setpoint and then have a good indication if it trips. A bright LED with a relay click for example.

As you probably have read here, there are good reasons to look for a good used bench supply from the great many reputable manufacturers.

If you can find the service manual for it online and can see it uses standard parts thats a big plus.

Then it is likely to be fairly easy to fix if it breaks.

[6PTsocket]

I am starting to see what you mean. Reviewers find things like big voltage spike when shutting off that fries sensitive electronics. As for Mastech, just another Chinese company building to a low price.
These things can have problems I never anticipated. I'll look for better reviews.


Sent from my SM-G900V using Tapatalk

[ArthurDent]

For a general use power supply I don't believe you really need a expensive precision power supply but one that is stable and doesn't drift is better. I have also found if you could have just one power supply that a dual or triple output supply is extremely versatile. Although I have several with digital meters I find that if I'm trying to read the current drawn by a circuit that an analog meter is easier and quicker for me to understand. I do have one single output power supply that has digital meters with a bar graph type gauge under the digital indication to give me both types of display on the same meter.

I have had a Mastech power supply and wasn't impressed with the design or the quality. One supply I found that seemed to fit the bill and was very well made was a Topward TPS 4000 with dual 30V/3A supplies and a fixed 5V/3A output. This supply (and similar ones) are available on eBay for about $75 and up in good used condition. Being linear power supplies they are quite heavy so you may get hit on shipping depending on where it has to be shipped from. The one I bought was from a local seller so I could pick it up and eliminate shipping charges. Here is a photo of the Topward supply and a extremely capable single output HP supply I have that could cost 4 to 10 times as much. The price you pay depends on what your needs (wants) are.
 

[cdev]

The GW-Instek GPC-3020 that I have is very similar to your Topward in functionality and layout and cost me around $55, lower than its typical price due to the large number of them which became available after a private for-profit technical school in the US went bankrupt. Its a solid unit although the fan is pretty noisy, I still have not got around to replacing it.

[h4nd]

I have a MCH-305D. It all seemed OK, but when testing LED / zener voltage, it was killing them.

I would set a 100mA current, then 10V, then drop the device onto the output.

I measured the current on a 'scope, and found that the current feedback was slow, so for a moment the PSU was poking 6A into the device (current overshoot). This was in excess of the 340mA abs max for small Zener I was testing, so - pop.

It was Ok if I set up the PSU, then powered up; I assume the slow rise time gave the current limiting circuit enough time to operate.

Anyone know if the Topward like the TPS-4000 have this nasty habit?

[xavier60]

I have a MCH-305D. It all seemed OK, but when testing LED / zener voltage, it was killing them.

I would set a 100mA current, then 10V, then drop the device onto the output.

I measured the current on a 'scope, and found that the current feedback was slow, so for a moment the PSU was poking 6A into the device (current overshoot). This was in excess of the 340mA abs max for small Zener I was testing, so - pop.

It was Ok if I set up the PSU, then powered up; I assume the slow rise time gave the current limiting circuit enough time to operate.

Anyone know if the Topward like the TPS-4000 have this nasty habit?

That's  common for most power supplies. My expensive Agilent U8002A psu outputs uncontrolled current for the first 100µs, mainly limited by the lead resistance.
Although my home made bench supplies have controlled current overshoot. The energy stored in the output capacitor still easily damages zenners and LEDs.

[bob91343]

It's possible that this surge is due to the output capacitor discharging.  It's a difficult design problem to go from minimal output impedance to very high output impedance.

For zener testing it's better to use a series resistor with a constant voltage source.  No way can you get a major current spike that way.  Current limiting is best used for preventing high load dissipation in the steady state, and transients can't be controlled that well.

[xavier60]

As well as for protection, my main uses for CC are for locating shorts on laptop main boards and for very accurately measuring shunt resistors while in circuit. Set  to 500mV and 1 amp, very unlikely to damage anything.

[MosherIV]

Hi

30V 5A - that is a difficult ask for a linear psu. It is the 5A that us the difficult part. Not impossible but it will make linear more exoensive.

I would NEVER trust the panel meters of any psu. I always double check with my dmms first.
So, my advice is to not worry about their accuracy or resolution, just put up with checking with a dmm first!

Set point interface. That is down to your preference, single pot or dual pot. Very few, in fact almost none, have multiturn.
You are right to make sure it is right for you. Again that is down to personnel prference.

You said that you do not mind linear or smps.
Consider what you want to  use it for?
If you want to do any work with op-amps or transistors, then a linear would be essential. This is because any psu noise will be seen in the circuit.
For everting else, microcontroller, audio, battery, digital, etc smps is ok

You have not set whuch country you are from.
In USA or Canada, you can get used HP easily on ebay.
In Europe you can get TTI (or older models from Thurlby) easily on ebay.
These are the 2 main brands of linear psu I trust.
I have an old Thurlby dual output and a HP, both obtained used and still perfectly ok

[tautech]

MCH K305D ............if you know how to use it. Had one for years and before I got some in for a customer we tried to bust one and couldn't !
The display drivers died a year or two back, fixed those and it's still doing the business of a basic PSU and some SLA charging.
Thread: https://www.eevblog.com/forum/testgear/cheap-variable-switching-power-supply/
Post #45 for the schematics.


Or you can go uptown for a decent 30V 5A SPD1305X linear PSU with extremely low ripple and noise that has sense capabilities too:
https://www.siglentamerica.com/power-supplies/spd1000x-series-programmable-dc-power-supply/

[Dabbot]

One supply I found that seemed to fit the bill and was very well made was a Topward TPS 4000 with dual 30V/3A supplies and a fixed 5V/3A output.

I have one exactly like this, but it is branded Parameters Model P4303. I wonder how many more rebrands of this power supply are kicking about, for those so interested in scoring one from eBay.

[Old Printer]

I have picked up a few of these little single output HP supplies. They are built like a tank, dead silent, manuals are readily available and can be found on ebay for about $50 in decent shape. They do have analog meters, but are accurate and the fine turn pots on the center knobs are very good. They are available in a number of volt/amp versions and can be chained together easily.

[MarkF]

I have picked up a few of these little single output HP supplies. They are built like a tank, dead silent, manuals are readily available and can be found on ebay for about $50 in decent shape. They do have analog meters, but are accurate and the fine turn pots on the center knobs are very good. They are available in a number of volt/amp versions and can be chained together easily.

+1.    I have four of them.  Only drawback is low current (500mA) capability.
I found a pdf version of the manual for download.

[rstofer]

Have you noticed that there isn't a sticky at the top of the page "Build This High End 30V 5A PSU"?

It's because it is darn difficult to do and most of the DIY circuits you find around the Internet aren't going to perform worth a darn.  This topic comes up every week or so, sometimes more often, and the results are the same.  Chase a design down a rabbit hole, concede that it doesn't work very well and ride off into the sunset.

No, I don't have any suggestions!  I bought a Rigol DP832 and it's good enough.  But it's an expensive supply and it seems unlikely that a DIY circuit is going to compete. 

How many of Dave's videos feature a DIY power supply when he is demonstrating some other kind of project?  He's a guy who could come up with a terrific design but, in the videos I have seen, he is using commercial supplies.  No, I haven't seen all of his videos.

The first problem is the specs:  5A is a LOT - Why so much?  Is it a real requirement or a random keystroke?  1A powers a lot of CMOS!  The same with 30V...  Why?  Sure, there may be a reason, some specialzed requirement, but by and large the specs come from copy and paste.

The problem isn't limited to 5A at 30V, the real problem is 5A at 1V in a 30V linear supply.  That's an enormous amount of heat in the pass transistors.  Figure an input voltage of, say, 36V.  That means the transistor(s) have to dissipate (36-1)V * 5A or 175 Watts.  That's why commercial power supplies have transformers with multiple secondary taps and relays to select an appropriate input voltage to the regulator for a chosen output voltage.  Getting such transformers may be a serious challenge.

In a lot of ways, quality in linear power supplies is proportional to weight.  Heavy supplies tend to be high quality, robust units.  The alternative is light weigh switching regulators.

Have you noticed there isn't a sticky "Buy This Cheap 30V 5A PSU From Alibaba! - Highly Recommended!"  Some of the switching supplies may be adequate.  Maybe they can be used as a preregulator to a linear supply.  But I don't see anybody jumping up and down shouting "Here is the answer to a maiden's prayer!".

I suspect that eBay may actually be the right answer.  Something from HP is going to outperform just about any DIY project.

[xavier60]

The design I used in my recent bench supply project works properly and would be safely capable of 30V 5A with tap switching and suitable heatsink.
Because of the size transformer that I chose, it safely outputs 4A from 0V to 15V and 2A from 15V to 30V by switching the 2 secondaries in parallel or series. Likely more due to the fan cooling of the transformer.
 https://www.eevblog.com/forum/projects/linear-lab-power-supply/msg2388873/#msg2388873