Tenma 72-8695 2 x 32v 3A + 5V @ 2A mini-review / teardown

[mariush]

Tenma 72-8695

http://ro.farnell.com/tenma/72-8695/power-supply-2-x-32v-3a-5v-2a/dp/1836057

TENMA - 72-8695 - POWER SUPPLY, 2 X 32V, 3A + 5V, 2A


    POWER SUPPLY, 2 X 32V, 3A + 5V, 2A
    Power Supply Output Type: Adjustable
    No. of Outputs: 3
    Output Voltage: 32V
    Output Voltage 2: 32V
    Output Voltage 3: 5V
    Output Current: 3A
    Output Current 2: 3A
    Output Current 3: 2A
    Length: 380mm
    Width: 600mm
    Height: 245mm
    Plug Type: UK / Euro
    Supply Voltage Typ: 220V AC

Price:  711 RON + VAT (about 130 UK pounds + VAT)

I was planning to buy a linear power supply for a long time and even started a thread here asking what should I buy.

I finally "bit the bullet" and bought this power supply from Farnell, and I chose this one for several reasons:

* the shipping from Farnell to my country is cheap (about 5$)
* being a EU seller, I would have less problems with warranty in case it's a dud

* it's a dual output power supply at just a bit extra compared to single output power supplies
* the outputs can be joined together to get a 64V 3A output

* Last but not least, this model looks very much like the Uni-T UTP3703

Now, I have a Uni-T UT61E multimeter which I like, so I was attracted to this model because of this, but just before I bought this one, I was told by a Uni-T importer that the company decided to stop working with the factory that produced their power supplies, due to inconsistencies in the build quality.

He told me that the supplies are fine but some of the issues he noticed were that some models didn't have proper grounding, or some buttons latched as long as the buttons don't get stuck, it's properly grounded, doesn't have wobbly feet or the outputs are sometimes split.

Well, I am pleasantly surprised by this power supply.

The power supply came with both a UK style power cable and a regular EU power cable and a small 4-5 page booklet which basically is the manual that's listed on Farnell's product page.

The power supply was powered ON (power button pressed inside the package) so it was probably tested before shipped or when it was received by Tenma from the factory.

Buttons, Knobs, Front

Led digits display, bright red... not too bright to be annoying.

The ground in the front is properly connected to the ground of the mains, and by default the metal link connects the ground and + of the secondary side. With this metal link attached, (as far as I can tell, I only had a few minutes to do basic tests) it's not possible to adjust the voltage of the second output, no matter if the power supply is in tracking mode or not.

When the Tracking mode is activated (button is pushed), the second output tracks the adjustment of the first. I'm not sure (didn't test) if the current limiting works when both are linked, or which current limit is respected (I assume the first output's current output limit)

The Measure buttons switch between showing the volts or amps for each output. The output needs to be shorted out to adjust the current limit  but a rough amps level can be set just by looking at the dot on the pot ... the Amps pot is not the multi turn kind.

The volts adjustment pots are 10k ten turn pots (or something like that), the Amps knobs are regular pots. The volts adjustment is pretty fine, can turn it a bit before going up or down 0.1v.

Volt Pots are made by... I don't know... there's load of them on Alibaba ... probably Vatronics  : http://www.vatronics.com/Upfiles/down/201171847057.pdf

Case, Feet

Feet are screwed in with thread lock, case is thick metal with nice big openings on the side for ventilation, the strap at the top is also attached with screws.

Mains, Grounding

Transformer, front jack, back connector properly grounded.
Transformer is tied down with some nice screws, I don't see it getting loose of vibrating.


Power transformer, power board

A toroidal transformer is used - I'm not sure of it's maximum current power. 

Two identical sections, each with:

* 4 diodes  1N5408 (3A max) : http://www.next.gr/components-datasheets/1N5400.pdf  They're spaced pretty well so I'm not concerned they may overheat the capacitors near them

* HA17324A quad op-amp : http://documentation.renesas.com/doc/products/linear/rej03d0674_ha17324a.pdf

* L7815CV 15v linear regulator for the op-amps probably

* 3300uF / 50v 105c capacitor for each side

* BU406 epitaxial silicon transistors :  http://www.fairchildsemi.com/ds/BU/BU406.pdf

* 2SC3281  NPN transistors (no-name on them) attached to the back heatsink with small to-220 heatsink in front of them  (assume these are for current regulation)

* Songle SRG-24 relays (four)  : http://songle.com/pdf/20085271545271000.pdf

* four adjust pots at the top for each side

There's an additional bridge rectifier on the right along with a capacitor but I'm not exactly sure what it's for... maybe it's for the front panel meters.
 
I couldn't see and I didn't check the actual output jacks. Other than seeing that the green jack is grounded to mains, I'm not sure if they have protection  (diodes) and if they're properly soldered there.
The buttons in the front seem to be good buttons with good switches and replacing them should be easy if they latch.

There's a whole bunch of wires on that board and I didn't want to mess with it now, didn't have time to go in detail on that small board, sorry.

Front display, Meters

Plain simple meter chip with led digits output

*  ICL7107CPLZ in PDIP 40 packages  : http://www.intersil.com/content/dam/Intersil/documents/fn30/fn3082.pdf

Should be easy to mod if someone wants, to get it to refresh faster (sacrificing accuracy) or anything else.

5v output

The 5v 2A output is performed by two L7805CV regulators linked together and screwed down on the bottom of the metal case. There's a 470uF 16 capacitor and a diode for protection on the jacks.

Other notes

It seems like a very simple design. I particularly like that:

* the boards that I see are not bodged.. I don't see anything obvious as fix or bodge.

* all wires are soldered to the boards - I don't see connectors with glue on them, no risk of cables being detached during shipping . And they're all tied nicely at various intervals.

* all capacitors are Chang ( http://www.huaweiec.cn/index_en.asp ) - while not a brand name, they're not exactly huang lo, crap no-name or rebadged caps.... they're owned by Huawei and I've seen them in various equipment...  basically it's refreshing to see they didn't just grab a bags of capacitors and threw them on the boards.

* some of the chips are brand name (or very good fakes) and they didn't bother erasing what's written on the chips.. makes it easy to fix it if something happens.

* lots of heatshrink where it should be.

* Power cables are nice, thick, and they provided both UK and EU cables.


Not liking

The only flaws/ not so nice that I see on a first look is :

* the transformer doesn't have power ratings on them

* not really a flaw, but some of the wires going into the boards have a bit too much un-insulated area. Seems like the plastic melted or shrunk due to heat when wires were soldered

Also, I'm not sure that I like that one of the NPN transistors is so close to the main power jack. It is properly insulated, with heat shrink and so on but still.

Other notes

* I would have liked to see 10 turn pots on the amps knobs
* When shorting the output, it seems the amp limit goes up to about 3.2 A

Pictures below :

If you want the hi-res (~ 3200x2500) download the pack here : http://savedonthe.net/a/tenma_72-8695.zip  (if the link goes dead in the future, pm me and I'll send them to you)

(edited the post to remove the part about fuse not used, it was incorrect. It is connected/fused properly)

[mariush]

Second set of pictures (exceeded 5 MB post limit with the first) :

[Monkeh]

!!!  I may be wrong, but the fuse in the power connector doesn't seem to be connected - the power supply may have been designed with the assumption that it's going to be sold in UK, and in that case the cable has its own fuse. See pictures... I'll double check when I have more time.

You're wrong. The fuse is used.

You do not use the fuse in the plug to protect the equipment. Please don't ever make that assumption.

I don't like the 5V output at all. Nasty little hack, and no option on the ground reference. But easy to improve.

[mariush]

I know the fuse on the UK power cable is not to protect device, but rather the house wiring and other electric equipment.. due to how the wiring is done in UK.

Like I said, I'll double check the fuse to see if it's used - I just saw the metal contacts unused (see picture) and assumed... I have some work now and can't do it right away.

What's wrong with the 5v output? It's just to 7805 connected together for 2A. Kind of hard to screw it up...

PS. Ah, nevermind. I see it now it is fused ... looking at the picture I can see where the contact with the cable is and where the actual cable inside goes, at the other end of the fuse. It's fused.

[Monkeh]

I know the fuse on the UK power cable is not to protect device, but rather the house wiring and other electric equipment.. due to how the wiring is done in UK.

It's to protect the cable, nothing else.

What's wrong with the 5v output? It's just to 7805 connected together for 2A. Kind of hard to screw it up...

Ugly implementation, forced earth reference, and the packages are subject to stress from flexing of the chassis (which is not meant to have ICs secured to it!).

[saturation]

Thanks for making these.  It appears decently made, what concerns are brought up about construction and parts placement are not critical.  Now the bigger question is how does it work?  Chances are it will be as decent as its build quality, but always best to find out first hand.

[T4P]

"Tenma" is just a rebadging operation from farnell, otherwise why do you think they sell UNI-T multimeters in orange 
Anyway the entire circuit reminds me of something of mine ... pretty sure they copied this one (this was 33 years AGO!)
Lots of quality difference and this was 1980's "cheap shit"  (The output transistors are not attached to the chassis, it's attached to a secondary PCB covered by that screwed in aluminum cover at the back of the heatsinks

[mariush]

I did a few small tests on the power supply the past hour or so.

The voltage meter is very accurate.. well, as much as I can tell compared with my uni-t UT61E meter ... I don't have any reference voltage at the moment to check my meter but I don't have any reason to assume it's not accurate.
It tracks very well, up to about 32.1v - after that for example 32.4v is showing 32.36v on the meter, but I guess it's close enough, the psu is 32v max. Also, the 22k meter already is a few counts off at that voltage.

The amps I only measured at low values - for example it was saying 0.15A and my meter on the mA scale was showing 144 mA. Close enough.

Tried the current limiting and it works, but it does it by limiting the voltage.  Put a 12v with 0.18 A rating on it and on a high limit (so not triggered) it used 144-150 mA ... when I lowered the amps limit to 0.10-0.13 A the voltage went down to 9v, then as I moved down closer to 0.10A, it went as low as 5v.

Below 0.1A, it's really hard to adjust. I'd say at about 0.07-0.08 A, the relay inside triggers and decouples the power supply.

I may need a refresher on oscilloscope usage but I simply couldn't trigger the oscilloscope to see if there's overshoot when powering the psu. I've set it to 5v and the scope and my meter showed 5.08v, turned it off and on several times and I couldn't trigger anything. 

Have to watch that nice long video about oscilloscopes that was posted here... it's queued in my playlist for weeks already. I have to admit I  don't have much experience with scopes so it's possible I screwed something up.

The ripple was about 10-20mv which I think is OK considering i just had the probe near the power supply and the meter and there was probably interferences and other factors.

The psu switches the relays when going up to about 18-19v, and switches the relays again when going down to about 16.5-17v.

The 5v fixed output is steady at 4.997v.

It kinda sucks to have to short the outputs and switch to measure Amps just to see the actual current limit. The current limit pots are kinda rough, you move it a bit and it jumps up by about 0.3 A. I may mod it to put 10 turn pots there.

[Bored@Work]

Tried the current limiting and it works, but it does it by limiting the voltage.

Ehm, voltage and current are linked to each other by something called Ohm's Law. What else than the voltage going down do you expect to happen when you have a constant load?

[mariush]

Well, I assumed that if I limit the fan to 0.1 A when it normally uses 0.15A, the power supply would simply remain at 12v 0.1A and the fan would spin slower... like how it is if I were to connect the fan to a bunch of batteries that are almost depleted... the batteries would still show 12v yet unable to give a lot of  current.

I know about ohm's law.

[Bored@Work]

I know about ohm's law.

Then apply it.

[ejeffrey]

Well, I assumed that if I limit the fan to 0.1 A when it normally uses 0.15A, the power supply would simply remain at 12v 0.1A and the fan would spin slower... like how it is if I were to connect the fan to a bunch of batteries that are almost depleted... the batteries would still show 12v yet unable to give a lot of  current.

Not to be a dick, but you just have a very basic misunderstanding of current and voltage here.  Pretend the "fan" is just a resistor of 80 ohm (it isn't, but that is close enough to demonstrate this point), and try to figure out what happens when you connect a power supply or battery with a 100 milliamp current limit.

[mariush]

I understand now. I = V / R 

If the fan is the constant resistance then the only way the current can be limited is by reducing the voltage.

Thanks for reminding me.

[T4P]

Below 0.1A, it's really hard to adjust. I'd say at about 0.07-0.08 A, the relay inside triggers and decouples the power supply.

That's not good. A power supply should CC down to 0v and not just suddenly decouple the PSU

[mariush]

Well, it's a regular 10k potentiometer, so it probably just jumps from 0 ohm to what would be 0.08 A . You know the way regular pots work, I'd have to be very, very precise in adjusting - about 300 degrees are equivalent to 3.4 A

If I  replace it with a 5 turn or a 10 turn pot, it will probably work just fine. I might just do that when i place the next farnell order in a week or so.

[T4P]

Yeah but these modern cheap shit are not worth at all trying to get more accuracy, if it suddenly "decouples" the PSU
My 80's "cheap shit" supposedly is a thousand times better than current cheap shit
Digsys has my PSU too, just labelled 5A (mine does 4.31Amps at 30V even labelled as 3amps, go figure!)

[mariush]

Maybe I wasn't too clear... by decoupling I mean the relay inside disconnects so that side of the power supply is outputting 0v when I turn the pot all the way to lowest value.

It sort of makes sense to me, if I adjust it to 0 A, you wouldn't get anything out, right? So why keep the relay enabled?

10$ for a 5-10 turn pot doesn't seem too much extra if I'd be able to adjust by 0.05 A instead of 0.2-0.4 A jumps.

[T4P]

Maybe I wasn't too clear... by decoupling I mean the relay inside disconnects so that side of the power supply is outputting 0v when I turn the pot all the way to lowest value.

It sort of makes sense to me, if I adjust it to 0 A, you wouldn't get anything out, right? So why keep the relay enabled?

10$ for a 5-10 turn pot doesn't seem too much extra if I'd be able to adjust by 0.05 A instead of 0.2-0.4 A jumps.

Oh. In my PSU the relay is only for tap switching, i thought you meant your relay decouples the PSU below 80mA, i was gonna replace the pots in mine, it's getting very jumpy and it's one-turn after all, maybe change all of them to ten turn pots so i get extremely fine adjustment

[Daz]

Hi, I know this is a fairly old topic but as somebody looking into buying their first bench PSU I'd really like to understand something:

The 5v 2amp channel of this PSU (channel 3) has its negative terminal connected to chassis earth (the manual for this PSU that I found online also confirms this). As an electronics newbie, my understanding is that I can't use an oscilloscope on a circuit that has a mains earth connection. Is that correct? If so, then I guess the 5v 2A channel would be useless to me if I needed to use a scope on my circuit?

If all of the above is correct then it's a bit annoying that this PSU has the 5v terminals marked as red and black, the same as the floating outputs on channels 1 and 2. I guess they should really be red and green to signify positive and ground.

Do all the other brands of cheap 3 channel PSUs have a similar channel 3 setup, i.e. using 5v and earth?

Thanks,
Daz.

[mariush]

Technically, you could just plug the power supply in a socket that has no ground lead. Practically, you shouldn't.

You could also just buy a couple of heatsinks (they're about 2-3$ each),

http://uk.farnell.com/multicomp/mc33279/heatsink-to220-x-2-7-1-c-w/dp/1710624
http://uk.farnell.com/jsp/search/productdetail.jsp?sku=1710623
http://uk.farnell.com/abl-heatsinks/ls85/heat-sink-to-220-to-3p-7-1-c-w/dp/595810
http://uk.farnell.com/jsp/search/productdetail.jsp?sku=1703170

unscrew the regulators from the case,  screw heatsinks on the regulators and then have the regulators sit on something that's not grounded.  (Ex. get a small square of prototyping board, drill some holes on the corners of the protoboard and use some nylon spacers or something, solder the regulators on the protoboard.)

The case isn't sealed with anything, so you won't void any warranty doing this.

[Daz]

Thanks. So the regulators are only using the chassis as a heatsink... that's interesting, I understand now.

There are a few other brands of 3 channel PSU in the same price range... I wonder if their fixed 5v channels also have -ve connected to chassis ground? This is the kind of info that's not shown in the spec!

Am I right about not using a scope on channel 3 of this PSU if -ve is connected to chassis earth?

Thanks.

[Geoffa]

Hi ,

Does anyone have a circuit giagram for Tenma 72-8695 with the PCB 20100521 ?
Will be much appreciated to repair - otherwise excellent PSU.