Mastech HY3005F-3 Power Supply Teardown

[jamesp15]

This is my "first" tear down of a piece of test/lab equipment, I thought I would share with you all. 
I haven't seen any detailed tear down pics of this power supply and thought others might be interested.

I have had this supply for a few months now, and finally got around to checking out its "guts". I am really pleased with its performance, low to no overshoot and its pretty much spot-on voltage and current settings (checked with  Fluke 87v, 289 and an Agilent U1272a).

Only a few (minor?) issues.

(Sorry I dont know how to get the picture "inline" on this forum.)

Picture 1 and 2 are the Front/Side Views

Picture 3 Left side circuit board, there is a "mirror" on the other side.  Note the hard to see in this pic crushed capacitor, and poor placement of components. (better images lower down in this post.)  Heatsinked TO220 is a 7812, and the non heatsinked one is a 7805)

Picture 4 and 5 Left side circuit board again, in 5 you can really see the crushed capacitor!

Picture 6 my 8 year old niece can place components neater than that!

Picture 7 overview, big toroid.

Picture 8 Right side circuit board

Picture 9 this appears to be the dedicated 5v output, 2 white wires come from the toroid, the red/black wires go to the dedicated 5v binding posts on the front panel, chip on this board is a LM723

Picture 10 back side of the front panel

Picture 11 more poor component placement.

Picture 12 inside of back panel, the transistors on the outside of the heatsinks plug into these boards.

Picture 13 back panel transistors.

If you want full sized original 18 megapixel images, you can see them here: http://www.bort.us/eev/hy3005f3/
(still uploading as of 19:10 GMT)

If anyone wants any more detailed pics of any part of it, or pics of something I looked over, let me know and I will post them.

[T4P]

Poor component placing aside, Mastech makes good PSU's that doesn't cost a bomb or two.

[madires]

Thanks for the tear down! I got a PeakTech 6075 which is nearly the same unit (up/down keys instead of knobs). It's a rebadged Mastech but I never saw a Mastech PSU with the same digital control. There's also the same poor component placement and some DC cables were put in parallel with AC cables. That was esay to fix :-) And about the filter cap I thought first they shortened the leads too much. Since 4700uF for 5A is on the cheap side I'll change that too.   

[saturation]

Thanks for making it available, and at good resolution.

I concur with DaveXRT, poor component placement and that crushed capacitor as minuses, but its less critical than some PSU we've seen torn down on eevblog. Its unlikely to be a performance or glowing poor safety issue.  My Mastech also had the issue of the AC and DC lines too close together, so I zipped tied them apart.

What I value more is the quality by which the electrical connections of the board were made, note this technology is circa early 1980s, late 1970s style, and is much easier to fix or troubleshoot.  Some observations:

The absence of gobs of glue you see in other PSU, often the kinds that don't work as electrically well?  It suggests they are confident about the integrity of electrical connections, solder or the snap fit, to hold properly, then don't even both glue things down as a just-in-case.

What's visible of solder joints suggest its all hand soldered and well done, almost no residual flux and cleaned.

Untrimmed long leads in the display board are generally straight.  Leaving it untrimmed puts it at risk of shorting if it bends too much, yet almost none are bent.  So some care was taken in assembly.

The parts placement has generous gaps and longer lengths of exposed leads, and lack of tight spacing makes it easier to clip test probes if needed, but there is no sign of rework, so it worked fine from the get go.

[bdivi]

Are those voltage and current pots multiturn ? If not fine adjustments would be imposible

[T4P]

Are those voltage and current pots multiturn ? If not fine adjustments would be imposible

Fine, maybe not. But you can just pop a multiturn inside it and just get away paying much less than a crappy rebranded PSU at farnell (READ:"Tenma")

[saturation]

DaveXRT is right, to make a good multiturn aka Bournes types, requires better quality and assembly control.  If you wish to add your own, its easy.  Ordinary Chinese production can easily make a good quality c 1960s design wire wound or carbon pot /trimmer and do it cheaply.  But I'm not sure I'll pay good money for a Chinese precision multiturn pot.

A single quality Bournes is about $30-40 each, retail.  5-6 of those alone cost as much as this whole PSU.

You can add an extra fine adjustment by making it a 3 pot adjustment instead of 2 or a single multiturn: just cascade 3 pots.  However, the real trick is the tempco stability of these pots and the overall design; a real Bournes is so stable you can extract uV resolution such as demonstrated in the 1960-70s design Power Designs precision DC PSU [ see archived thread]; also if the design can't regulate stably below 100mV or so, adjusting for such a value also becomes moot.

Are those voltage and current pots multiturn ? If not fine adjustments would be imposible

Fine, maybe not. But you can just pop a multiturn inside it and just get away paying much less than a crappy rebranded PSU at farnell (READ:"Tenma")

[Bored@Work]

Could the boards originally be designed for another PSU? It looks as if they stuck two boards, originally designed for a single output PSU (plus a fixed voltage), into this to make a dual output, plus fixed voltage.

[saturation]

For comparison here's a link to old photos from my single VDC output Mastech:

https://www.eevblog.com/forum/product-reviews-photos-and-discussion/lw-dqgs-power-supply-review-and-teardown-cautionsafety-risks!!!/msg58005/#msg58005

As you can see, the build quality is even better, but very likely the variation is due to manual assembly and not the strictest quality control.

[kampianakis]

Hello blog

I just received my Tekpower TP3005D-3. I bought it after I saw the stellar reviews online. I need it to power this big guy

http://www.minicircuits.com/pdfs/ZHL-16W-43+.pdf

And also to have a nice high power and relatively low-cost DC power supply. I picked this one mainly because I thought that the big toroid would provide some noise immunity compared to any kind of simple switching powersupplies (e.g. for notebooks). However, once I plugged it in, loaded it with a bunch of 100 Ohms  in parallel (I didn't have any high power resistors so used 1/4 instead and warmed up the place a little bit). In loading conditions the supply seems to regulate pretty well by inspecting both the oscilloscope and the multimeter.

However I checked the FFT on my Agilent DSO-x 3014A and it showed a LOT of spurious signals approx. 15dBs above the noise floor.
A few of the frequencies are: 127kHz, 7kHz, 450Hz. Some of them are probably alliased into the spectrum but in any case there are a bunch of them and this doesn't look good.

The application that I am working on right now is a bioamplifier and as you can imagine I can't afford any unwanted crap so that I can have repeatable experiments.

Oh I also compared the spectrum with an Agilent E3631A and I saw no problem with the Agilent :/

I have attached some pics with my oscilloscope output for you guys to check.

Long story short:
The power supply has spurious stuff. Has anyone observed something like that? Should I care too much?

[saturation]

Insure the scope isn't earth grounded and the PSU output is truly floating.  Otherwise, you'll pick up ground loop noise.   These old PSU designs are all analog so there really isn't a source to make such spurs unless is picking up EMI internally, e.g. digital panel meters or externally.

For details see how to make a PARD measurement.

http://www.imcpower.com/prodigit/an_403.htm

Hello blog

I just received my Tekpower TP3005D-3. I bought it after I saw the stellar reviews online. I need it to power this big guy

http://www.minicircuits.com/pdfs/ZHL-16W-43+.pdf

And also to have a nice high power and relatively low-cost DC power supply. I picked this one mainly because I thought that the big toroid would provide some noise immunity compared to any kind of simple switching powersupplies (e.g. for notebooks). However, once I plugged it in, loaded it with a bunch of 100 Ohms  in parallel (I didn't have any high power resistors so used 1/4 instead and warmed up the place a little bit). In loading conditions the supply seems to regulate pretty well by inspecting both the oscilloscope and the multimeter.

However I checked the FFT on my Agilent DSO-x 3014A and it showed a LOT of spurious signals approx. 15dBs above the noise floor.
A few of the frequencies are: 127kHz, 7kHz, 450Hz. Some of them are probably alliased into the spectrum but in any case there are a bunch of them and this doesn't look good.

The application that I am working on right now is a bioamplifier and as you can imagine I can't afford any unwanted crap so that I can have repeatable experiments.

Oh I also compared the spectrum with an Agilent E3631A and I saw no problem with the Agilent :/

I have attached some pics with my oscilloscope output for you guys to check.

Long story short:
The power supply has spurious stuff. Has anyone observed something like that? Should I care too much?


/quote]

[KG7AMV]

Anyone have a display schematic?

Looks like I may bee in for a display repair..

Found it.. https://www.eevblog.com/forum/testgear/mastech-hy3005f-3-seriesparallel-switches-imply-non-floating-supply/msg492343/#msg492343

https://youtu.be/TORUWRbDTbc

I am getting way to curious.. Bum seller or shipping damage? Shall the screwdriver tell the tell??

Yep.. Looks Like Smoke Got Out Some time Ago!!

[KHuffman35]

Solder quality on these units appears to be a significant issue.  I purchased mine about 7 years ago and use it sporadically for electronics work.  I don't know if the fixed 5V output ever worked, but discovered it was dead a few weeks ago.  Also, last year the master channel began to wander on the setpoint voltage.  I recently repaired both:

Fixed 5V output - A power FET is mounted against the back of the case with a heatsink.  The leads on the FET are attached directly to wires that route back to the fixed 5V regulator board.  Two of the three wire leads had fallen off the FET.  Thankfully, the wires were covered with heat shrink, so they didn't short against the case.  I reattached the wires to the FET and the fixed 5V works fine.

The master channel output, starting approximately a year ago, would no longer hold the setpoint voltage.  It was more pronounced at lower setttings and would wander by as much as 3 or 4 Volts off the setpoint.  I removed the control board with the adjustment pots (mounts directly behind the faceplate; held in place by nuts on the adjustment pots; pull knobs and remove the nuts).  Note - this isn't the power board mounted on the side rails in the case, but the board that connects to the knobs used to adjust current and voltage.   I reflowed the voltage & current adjustment pots as well as other connections on the master channel (only a few on this board).  The master channel is working as expected and is rock solid on the setpoint.

[Mikecast]

Thanks for the post on wandering voltages.  Mine is doing the same and I will follow your lead to see if that fixes my issues.

[dcarolan]

I just discovered this on mine too. Glad I saw this post I was overthinking what to do about it.