Does a hobbyist need a bench power supply?

[cybergibbons]

So, all these projects we are making all require power, so an essential piece of equipment is something to power all of these.

My first power supply was an Irwin variac power supply, which provided AC and smoothed, unregulated DC. I got it from a skip at school as the plastic casing was smashed and the banana plug sockets damaged. I then found it also had a 110V winding which meant you could get twice as much output voltage. It served me well until my dad chucked it out in a clear up whilst at uni.

Since then I have relied on wall-warts (for most digital/microprocessor based work), custom built (e.g. +/- supplies for amplifiers), and a single-ended programmable switch mode bench supply (a GWInstek PSP-2010  http://www.gwinstek.com/en/product/productdetail.aspx?pid=38&mid=75&id=164 which I got because it was £100 in clearance at Rapid, and is very useful for battery charging, testing motors and so on).

I'm feeling the need to get a proper bench power supply now - but I'm not quite sure why. Maybe it's because I see them on everyone else's bench. Maybe I'm just annoyed at having to build a 1.2/3.3/5/12/15V power supply each time I need one.

So what to people feel they need in a bench power supply? Is dual output with independent/parallel/series enough? Do we even really need a variable power supply for most projects? How much current is enough?

I've been looking at these two:
http://cpc.farnell.com/caltek/psm3-5a/power-supply-bench-led-3ch-30v/dp/IN05511 - which is mentioned here: http://www.rcgroups.com/forums/showthread.php?t=1030827
http://aspen-instek.com/01-multi-output-dc-ps.html (GSP-4303 - I've bought from these guys before, their prices are very good).

Any thoughts? Any bargains knocking around?

[alm]

In my opinion, yes. It's used to power something before the power supply is finished, and to independently test the modules. Or just for experiments.

I use the variable current limit to set the current limit just barely above the normal current draw, so nothing blows up when shorted. The variable voltage is useful to test things like minimum voltage (eg. dropout voltage of a regulator). There have been various topics on this forum about them. You can also build one yourself, they're not that complex.

Tracking is useful for op-amps which need symmetrical power supplies. Amount of current depends on your projects, I rarely use more than a few hundred milliamps, but I also know people that have and use 10A+ power supplies.

[Zed]

I would say that a variable voltage power supply is one of the things you need after a multimeter but don't bother buying one, build your own. My bench power supply is based on a LM317 variable voltage regulator, will supply current up to 1.5 amps but it only has a 1 amp current meter and best of all the voltage is variable from 1.2 volts to 13 volts which is pretty much all you need for doing hobby electronics. If you need current limiting, which is handy, you can use a L200 voltage regulator which will go up to 2 amps but the voltage will only go down to 2.85 volts. My next power supply project will prolly include + and - fixed regulators of various voltages for op amp work as well as the variable supply. The internet is awash with schematics and data sheets so you can easily find a design to suit your needs.

[charliex]

i'd say fairly essential, primarily for the reasons before, variable and current limiting.

less useful is the atx to bench conversion, since its neither of those
http://www.wikihow.com/Convert-a-Computer-ATX-Power-Supply-to-a-Lab-Power-Supply

[XynxNet]

For me a bench power supply is essential. I don't need variable current, but variable voltage is a must have.

I don't think it would be a good idea for a beginner to build one himself. Mains is just too dangerous, if you don't really know what you are doing. Fortunatly simple bench power supplies arn't that expensive anymore.

[Kiriakos-GR]

I need it all ... variable current ... variable voltage  10A to 30V, so to be able to test anything .

My advice , get the best , its the only tool that you will own it for a life time..
The ones with variable current are immortal    

[PeterG]

I use 2 power supplies. The main ps is a standard 0-30V 0-5A setup. The other is a 9-15V 30A unit.

I would be lost without either of them.

Regards
Peter

[MrPlacid]

As a hobbyist myself, I can see the vital importance of a bench power supply. It just makes life simpler and testing ideas much quicker. I just bought my first 0-18V 5A power supply last month and I am already searching for one with higher voltage of 50V and up. Not only can I run test circuits of it, I can also run other consumer equipments too.

Depending on what you are using it for, you'll require different voltages. For auto 12v, phone 48v, airplane parts 28v, etc..

The bench power supply has so much uses. We can even uses it to blow up capacitors (and fuses) just like Dave

[NiHaoMike]

You can almost never have too many power supplies, and at least one or two should be adjustable with current limit. Fortunately, they're extremely easy to build, especially with all the buck converter chips out there.

[cksa]

What are some popular buck converters with adjustable current limit?

[cybergibbons]

To all those that feel the self-built is the best route to go - can a home built supply even compete on cost with a bought one? There's quite a few high value parts in there:
* Transformer
* Case
* Smoothing caps
* Heatsinks
* Panel meters

I just remember the first audio amp I built from a kit (I think it was 150W or so)- the kit itself cost about £50, built and tested at school using a bench power supply. Then I came to casing it, and found that mains lead, fuse, power switch, grommets, terminals, phones all cost a fair bit. Then there was the power supply... cost an absolute fortune!

[saturation]

I couldn't agree more.  Unless your parts are salvaged from the trash, building won't be cost effective if premade assembled or a kit exists.  You can buy cheaply if you get a brand like Mastech.  At the very least, you get the hardware you can modify to improve the specs or repair, if necessary.  

The reason is economy of scale, inventory and distribution chain.  DIY is a 'custom build.'  By the time DIY new parts you need reach your distributor, and finally you, its gone through several hands that up the price, not to mention shipping, handing, taxes or duties.  

Engineers dedicate careers designing power supplies especially in this Energy Star conscious world.  Why reinvent a well established wheel except for the fun of it?

That said when I consider the simple general specs:

linear or switching
one, two or > 3 outputs
programmable

The unit you already have is more than necessary for most home lab work, but do you need more than one ouput?  With the ever increasing trend to low voltage, low power usage in devices, especially made to be portable and battery powered, a clear need is one with tight dependable regulation, and constant fully adjustable V and I.

To all those that feel the self-built is the best route to go - can a home built supply even compete on cost with a bought one? There's quite a few high value parts in there:
* Transformer
* Case
* Smoothing caps
* Heatsinks
* Panel meters

I just remember the first audio amp I built from a kit (I think it was 150W or so)- the kit itself cost about £50, built and tested at school using a bench power supply. Then I came to casing it, and found that mains lead, fuse, power switch, grommets, terminals, phones all cost a fair bit. Then there was the power supply... cost an absolute fortune!

[cksa]

Well you could just get a cheap not adjustable power supply, say that supplies 5 amps at 12V. Then you do all your DIY stuff after that. Shouldn't be too expensive?

[alm]

You can often get some of the parts cheap, like transformers, heat sinks and a case. If this is the case, DIY might be cheaper. Or if you want more advanced features, since commercial power supplies with features like GPIB/RS-232/LAN control get expensive fast, but it might only add $15 or so to your build costs. If you pay full retail, everything gets expensive.

Cheap lab supplies aren't usually very energy efficient, just simple linear supplies with at best a switchable second transformer winding, so I doubt that the designers spent a lot of time on that. When set to 10% of the max. output voltage, 90% of the power will be converted to heat, regardless of what you do.

I agree that high currents and high voltages are a lot less important for most modern devices. It's more important that you can accurately output something like 1.2 or 0.8V than 2x30V. But it depends on your applications.

[Rhythmtech]

I added an LM317 to my ATX conversion bench top.  I can get variable output from 2V - 12V and from 12V - ~23V.  The amount of current I can use depends on the voltage drop across the regulator, but I usually get the 300 - 500 mA I need for most small projects. I have a large adjustable supply from Sorensen if I need more, but I rarely do. The small size of the ATX power supply in my work area, the fact it didn't go in the trash and that I built it make it a tool I reach for often.  

I have been considering adding a buck boost regulator to raise voltage out on it, but only for fun since the majority of my projects are all TTL level.  I did a cursory check on digikey and came up with an AP34063, which for 76 cents did not seem bad, but would require a bit of work since it seems to rely on capacitance to change switching frequency to adjust the output. It would be nice to have a switched output instead of the linear also, so as to not worry about frying the regulator as I sometimes get close to doing on the LM317 when I run it at 3.3V.  So I am still on the lookout for a cheap buck boost solution. I also would like add a voltage readout(s) eventually.  

[saturation]

http://mastechpowersupply.com/dc-power-supply/linear-power-supply/mastech-regulated-variable-dc-power-supply-gps-3030d-30v-3a/prod_62.html

Just checking, they have over 50% discount on their PS.  This wasn't there in the winter, when I first took a look at their linear and switched supplies.

[slburris]

So their sale price is $90 for a 0-30V 0-3amp supply.
For about $170, what do you think about their 3 output supply:

http://mastechpowersupply.com/dc-power-supply/linear-power-supply/mastech-linear-dc-power-supply-30v-3a-hy3003f-3-triple-outputs/prod_7.html

Looks like two of the $90 supplies put together with a 5v supply thrown in
for less than 2X the cost.

Scott

[Rhythmtech]

I can't seem to recall what the requirements of the power supply we are supposed to recommend are?  Personally I kind of like the programmable one.  It's small so it won't take up too much space, allows you a measure of automation via a PC as well as apparently an arbitrary waveform generator. 

Upon looking for any information regarding the interface via manuals etc...  I came to find that, none were available through mastech's site for any of the DC power supplies.  At this point I would pay more to get a product with some support other than a pamphlet of unknown quality.  The support button takes you to a link to their forum, where the only thing you find regarding the programmable power supply is posts telling everybody it is new. 

Do yourself a favor and get a PS that at least has manuals posted online so you can check out how crappy the instruction are in advance and get a feel for how important the company feels you are. I had an Extech EX330 meter that broke and was shipped a replacement within a week, their customer service was excellent in my experience.

Check this one out - Extech 382213 $199.99

Here's Extech's page with manuals and a data sheet Way better than mastech's "support forum".

[cybergibbons]

Interesting discussion!

I agree that my programmable 20V/10A supply is enough most of the time. It's switch mode so quite efficient. I certainly couldn't use it as a arbitrary waveform generator though, as RS232 is just too slow.

I'm also a big fan of manuals and support, which I guess is why I have a GWInstek scope and not a Rigol... but then I could have bought two scopes for the same price.

That said, power supplies are a lot more simple.

[slburris]

I'm glad I'm not the only one having trouble finding Mastech manuals online.
But seriously, a Quakko power supply?  I'm still smiling at the name.  Aren't
they another brand of Aoyue?  I've heard of them only in the context of
SMD rework stations.

Personally, I tend to prefer the LED displays over the LCD -- easier
for my eyes to read :-)  If money were no object, I like the
Hameg 7044: http://www.hameg.com/163.0.html

But we're outside of what most hobbyists need, including me!

I suppose it's too much to ask for a power supply manual to contain
schematics these days (the Extech manual does not).


Scott

[alm]

Get a used one (eg. HP) from ebay . You'll probably be able to find a schematic for those, certainly brand names like HP. Except for switchers and digital control, I'm not aware of many improvements in bench supplies in the last fourty years or so, so an old one with analog panel meters might work just as well (I actually prefer analog panel meters, they have a much higher bandwidth and I don't expect accuracy anyway).

[cybergibbons]

I do like the way that HP test equipment, no matter how complex, always had the schematic. It amazes me what they managed to do without microprocessors.

Hakko is another brand that sounds like Quakko... maybe this is the first case of a clone of a clone company?

I agree with the LED over LCD - my current supply has an LCD display which can't be read from above at all, which is really irritating.

For some reason the UK has a very poor second hand supply of power supplies on ebay. Most of my finds have been from skips... I used to have several lovely valve scopes as a result.

[allanw]

I bought a used TTI PL330QMD dual output 30V/3A supply off ebay for only $50! The only downside was that it was wired for UK power, but the manual actually gave steps to convert it.

The current limit on one of the outputs was abnormally low, so I asked the manufacturer for a service manual and they actually gave me one. Turns out the problem was just some burnt out resistors, which I'll be replacing soon.

[Rhythmtech]

I do like the way that HP test equipment, no matter how complex, always had the schematic. It amazes me what they managed to do without microprocessors.

Hakko is another brand that sounds like Quakko... maybe this is the first case of a clone of a clone company?

I agree with the LED over LCD - my current supply has an LCD display which can't be read from above at all, which is really irritating.

For some reason the UK has a very poor second hand supply of power supplies on ebay. Most of my finds have been from skips... I used to have several lovely valve scopes as a result.

Not the first case for hakko -  Luckey 936

I've always had most of the ps's I've used at eye level mounted high on the bench, so I had not considered the LCD vs. LED issue. At home it would probably be much lower and I will definitely consider it from here on out.

I bought a used Sorensen industrial variable DC power supply from ebay and have had to do some repairs as it was in an industrial environment.  But I can't complain about getting a $1200 power supply for $30 shipped. It is not really a bench supply however, weighs about 35 lbs and is cumbersome because of its size on small benches. Eventually it may become part of a home test and measurement test board setup.

For my day to day small power needs I built an ATX conversion power supply out a 350W Antec power supply because I wanted a supply I could carry to meet ups, give to a budding techie if I wanted to at no great loss, or blow up without much concern. I had all the parts sitting around already and it was a good excuse to "clean up" productively. I would however prefer to eventually get a good bench supply, but my current setup allows me to more than get by so I am happy at the moment.

I like this - Extech 382280  too much $$$ for me though.

[saturation]

Yes, what's I see most work put in by designers is evolving PSU programmable modes, tighter parts integration, improved efficiencies and regulation.  These Mastech's have ~ volts and amp ratings, but the regulation is 2-10x worse on the SPSU.  At this price range, I'd be probably thinking of a better cost effective brand name, like Instek.  Consider too, the LPSU is 25#, while the linear is 10#.

http://www.mastechpowersupply.com/dc-power-supply/linear-power-supply/mastech-variable-linear-power-supply-30v-10a-hy3010d/prod_11.html

http://www.mastechpowersupply.com/dc-power-supply/switching-power-supply/mastech-power-supply-hy3010ex-30v-10a-over-voltage-over-current-protection/prod_70.html

So what does a small lab need?

More than anything else, I think an a priori in electronics is a PSU with clean stable output power. NiMH cells provide a low cost portable power supply, and typically output 1.2V at 2700mAH.  They need to be cutoff ~ 0.9V.  So, a PSU in the 0-15 3A works very well to emulate single cells to 14 NiMH cell packs, at least in my application.  Also, most analog and digital components easily work in this voltage range and power requirement.    

In this application, a modest linear can provide a home lab user an easy to repair, quality PSU so a brand name is less critical.  This smaller Mastech seems just right, but because this device is seems well suited doesn't make the whole Mastech lineup equally outstanding.  Mastech documentation isn't widely available on the net.  Some schematics are here:

http://www.eserviceinfo.com/equipment_mfg/Mastech_11.html


A simple example:

Mastech HY1503D is $50, 0-15V 0-3A


Mastech is also sold as the Kaito brand, and there have been some complaints in the high current output devices on Amazon.  I also would not trust its programmable models, given the need for good documentation and support in that area and would easy go form Instek.

You can often get some of the parts cheap, like transformers, heat sinks and a case. If this is the case, DIY might be cheaper. Or if you want more advanced features, since commercial power supplies with features like GPIB/RS-232/LAN control get expensive fast, but it might only add $15 or so to your build costs. If you pay full retail, everything gets expensive.

Cheap lab supplies aren't usually very energy efficient, just simple linear supplies with at best a switchable second transformer winding, so I doubt that the designers spent a lot of time on that. When set to 10% of the max. output voltage, 90% of the power will be converted to heat, regardless of what you do.

I agree that high currents and high voltages are a lot less important for most modern devices. It's more important that you can accurately output something like 1.2 or 0.8V than 2x30V. But it depends on your applications.