Author Topic: Why is BK Precision still in business? BK8500 programmable load OEM found  (Read 71571 times)

0 Members and 1 Guest are viewing this topic.

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #50 on: September 22, 2012, 09:34:08 pm »
Very nice report, LaurenceW it couldn't be more comprehensive.  It sounds like the unit is doing very well for its thermal management, and if you can't detect an imbalance in the drops across those sense resistors nor a temperature gradient then for practical purposes the temps and loads are ~ similar one end of the heatsink to the other.  This is very good news, so much rides on cooling when using eloads.


..I cannot detect any significant variation or changing voltage drops in the series 0.2R resistors even over several minutes of full load operation... electronics isn't actively watching the individual mosfet current (as a voltage drop across each mosfet's 0.2R series resistor), and doing some load balancing that way. I am not about to do a further tear down and track out the mosfet gate drive electronics. but I could see that theory being the case.

(Electronic) Loads of fun!
Best Wishes,

 Saturation
 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #51 on: September 25, 2012, 05:57:48 pm »
Hi LaurenceW, and follow up on the calibration issue?
Best Wishes,

 Saturation
 

Offline LaurenceW

  • Frequent Contributor
  • **
  • Posts: 258
  • Country: gb
    • It's Time, Jim, but not as we know it
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #52 on: September 25, 2012, 08:13:21 pm »
Hi Saturation, Still haven't cracked the calibration issue, no! I have at least had a reply from the Hong Kong retailer- a one line response to my (I thought my very detailed) explanation, and it makes no sense to me! I will try again, but I think I am on my own... :(

Here is the top and bottom of it.

Although the manual is slightly less than crystal on the subject (!), the Maynuo features two voltage ranges and two current ranges. These are not auto-ranging a-la multimeter. The lower ranges work over small voltages and currents, but give you more resolution. The ranges are

Voltage
LO: 0-20V with Four places of decimals displayed - and I think probably about believable. Given a stable voltage, these figures are stable; no noise or jitter
HI:  0-500V with Two places of decimals. Again, stable

Current
LO 0-3A with 5 decimal places shown, but the last one is iffy at best
HI 0-15A  with four believable decimal places

A relay can be heard to click when selecting between the two current ranges. This relay is not in the current path at all, but may be switching the point on the current shunt from when the voltage (representing the current) is taken - dunno for sure.

The Maynuo features NO adjustable pots inside - good. Instead, the instrument is calibrated from the outside, by a built in calibration routine, which simply requests that you feed the load certain defined voltages and currents.  You are invited to enter two pairs of voltages (which are used to set the hi and lo voltage calibration), and then two pairs of current, (allegedly) to set the hi and lo current calibration. I set all these against my Fluke 289 which I know to be pretty-well spot on, within it's own limits. Actually, the cal routine asks for certain value, like 2.0000V and 20.000, but you can just set your voltage source to something close like 2.0123V, and then key this actual value in. The software then interpolates this. That bit seems to work - nice.

So, I've got the voltage and low current calibration values spot on. But I cannot cal the hi current range. I notice that it is at least consistently wrong, so this evening i am about to do some deliberate "miss-calibration" tests and, like any good lab experiment, write the results down, and have a look.

I think I am looking at a BUG in the calibration routines, but I might be able to work out what current it is asking for, and work backwards...

**

GRR! I've also got a separate problem, in that I now cannot re-establish proper two-way comms with my PC. That's for another day...
If you don't measure, you don't get.
 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #53 on: September 26, 2012, 04:33:57 pm »
Thanks for the update, LaurenceW,  this is the Maynuo's only true visible weakspot.  I took the liberty to research it and can't locate calibration protocols on BKPrecision or Itech either.  It seems to be left out of the manuals or other support literature.

Maybe you can get better luck writing BK?  The register storage format for all 3 items is =~ so maybe asking BK by email will be =~ solution as well?

While the adjustable pots are laborious, they are foolproof, low tech and easy to do with just about any volt or current reference, so I trust the Chinese can do it that well. 

My experience with close case calibration is its better if well done and if flexibility is allowed with calibrators used.   Its a boon if you have automated calibrators, but it can be more work if you're winging it with lab substitutes.

Its impossible to self-cal a Fluke 87V for example, unless you have access to all ranges required and in the proper sequence [ easiest using Fluke's $30,000 calibrator ] and completed.  Agilent 1272a's calibration routine is exactly as written but you can do any range any time and is a joy to self cal using any reference, but an earlier model 1252a doesn't respond as expected: enter 1.000 mA for example and it sets itself as 0.999 mA, that kind of glitch,  and one has to manually offset it by using the delta, e.g. 1.001 mA to get it to register 1.000mA.  Sounds like your problem too.  And yes, the voltage ranges are easier to set and respond as expected.  Even if the reference has some impedance issue, the meter reads the right current, yet, enters it with an offset.  Its a PITA.

However, old pot style Array 3710a, there are 6 pots to adjust, 2x for each range, hi and lo,  then another 2 for meters, and WYSIWYG with the Array.

... , <snip>
A relay can be heard to click ...

The Maynuo features NO adjustable pots inside - good. Instead, the instrument is calibrated from the outside, by a built in calibration routine, which simply requests that you feed the load certain defined voltages and currents.  You are invited to enter two pairs of voltages (which are used to set the hi and lo voltage calibration), and then two pairs of current, (allegedly) to set the hi and lo current calibration. I set all these against my Fluke 289 which I know to be pretty-well spot on, within it's own limits. Actually, the cal routine asks for certain value, like 2.0000V and 20.000, but you can just set your voltage source to something close like 2.0123V, and then key this actual value in. The software then interpolates this. That bit seems to work - nice.

So, I've got the voltage and low current calibration values spot on. But I cannot cal the hi current range. I notice that it is at least consistently wrong, so this evening i am about to do some deliberate "miss-calibration" tests and, like any good lab experiment, write the results down, and have a look.

I think I am looking at a BUG in the calibration routines, but I might be able to work out what current it is asking for, and work backwards...

**

GRR! I've also got a separate problem, in that I now cannot re-establish proper two-way comms with my PC. That's for another day...
Best Wishes,

 Saturation
 

Offline LaurenceW

  • Frequent Contributor
  • **
  • Posts: 258
  • Country: gb
    • It's Time, Jim, but not as we know it
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #54 on: September 27, 2012, 08:05:47 pm »
Saturation, Looks like just you and me having a private natter on this one... anyway, here is today's update.

The GOOD NEWS is that emails directly to Maynuo are eliciting a response! Not fulsome answers, but just requests to recalibrate this and that (getting quite good at that, now!)

Constant current, power and resistance all works, but setting a constant voltage causes the Maynuo to clamp to about (and a very consistent) 85% of the V-Set value.  Photo shows load with Vset to 10V on a 15V 0.5A max power supply, but the Maynuo clamping this to 8.447V (8.499V says Fluke, but I'll give it the 2mV...)

So much for the Maynuo's "unique feature" over the competition! :(

We press on...
If you don't measure, you don't get.
 

Offline robrenz

  • Super Contributor
  • ***
  • Posts: 3035
  • Country: us
  • Real Machinist, Wannabe EE
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #55 on: September 27, 2012, 08:31:27 pm »
Saturation, Looks like just you and me having a private natter on this one... anyway, here is today's update.

Dont you believe it.  I am reading every word.  both of you please keep up the good work.

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #56 on: September 27, 2012, 08:39:38 pm »
Thanks LaurenceW, always good to read your reports.  FWIW you have the only detailed review on the Maynuo on the net so once its archived, at least others can read it for posterity.  Often only a few folks ends contributing to a discussion, as eloads aren't a high interest item like DMM or DSOs.  Good to read robrenz here, we had the same issue with the Power Designs supplies, just maybe 2 or 3 discussants but it went a long way.

Great, you're testing their customer support!
Vset is annoying, but if its consistent you could work with it?  i.e., adjust the Vset up past 10 V until the proper voltage shows on the input?

On the side, I presume your test 15V 0.5A PSU is capable of providing the ~7.5W, some less well PSU end up dropping Vout when pulling the rated amps.  You could check out the test PSU first, set the eload to CC, start at 0.1A and adjust up and see if the test PSU voltage stays a stable 15V as you dial up to 0.5A.

Maynuo still plays like a good machine, like any V1.0 it probably has a few firmware issues here and there if the test PSU turns out good.


Saturation, Looks like just you and me having a private natter on this one... anyway, here is today's update.

The GOOD NEWS is that emails directly to Maynuo are eliciting a response! Not fulsome answers, but just requests to recalibrate this and that (getting quite good at that, now!)

Constant current, power and resistance all works, but setting a constant voltage causes the Maynuo to of the V-Set value.  Photo shows load with Vset to 10V on a 15V 0.5A max power supply, but the Maynuo clamping this to 8.447V (8.499V says Fluke, but I'll give it the 2mV...)

So much for the Maynuo's "unique feature" over the competition! :(

We press on...
« Last Edit: September 27, 2012, 08:41:44 pm by saturation »
Best Wishes,

 Saturation
 

Offline LaurenceW

  • Frequent Contributor
  • **
  • Posts: 258
  • Country: gb
    • It's Time, Jim, but not as we know it
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #57 on: September 28, 2012, 01:51:09 pm »
Hello Robenz, glad to know are following along! It gets kinda lonely otherwise...

Saturation, you raise a valid point about the "manliness" of my power supply - but rest assured, no concerns with this puppy! see pictures.

It's a TTI Triple programmable power supply PL330TP, good for 2 x 32V @ 3A each, plus 4-6V (i.e usually 5V) DC at 7A. All independently floating with 3200-count volt and amp meters on the main supplies (down to 10mV and 1mA resolution and[/i] good for that). So I can get combinations of up to 70V DC at 3A or 5V DC at 13A! The two main supplies don't automatically track one another (for symmetrical +/- 15V op amp supplies and the like), but I've no real need for that. The supply is remotely programmable via RS232 (I've verified that it works, but I have no pressing need to use it).

This supply is built like the proverbial brick sh!thouse!  I will give folks a tour of it, one day. I bought it as new, but end of line. TTI have moved on to a new range of power supplies now, but from what I have seen the changes are largely cosmetic. This generation of TTI supplies do crop up on FleaBay from time to time, and by and large are a SOUND buy, IMHO. I have no doubt that this piece of gear will outlast me. I couldn't say that for ANY of the other Chinese power supplies I've seen tear-downs of, on this and other sites. This TTI weighs a tonne - Four transformers! FOUR! All linear stuff, massively overspec'd output stages. No fans - no need. Very well behaved. No SMD components, fully documented with circuit diagrams and all.

So, I think it's good enough for the 7.5W, (or up to 96W per Ch), Saturation :) :)

(two pics, one of the load set to 2 x 32V and 3A current limit, plus 5V. Other with flash, but short camera exposure time shows misleading display)

ANYWAY - where were we? I set aforementioned power supply to 15V at 0.5A Max, so that I could prove (whether or not) the Maynuo M9712B would securely clamp the supply to a required 10V. Alas, it does not, seeming to think that 8.4V is "good enough". Yes, I can get the load to clamp at 10V by requesting a V-set of 11.91V, but clearly something ain't right...
« Last Edit: September 28, 2012, 01:58:42 pm by LaurenceW »
If you don't measure, you don't get.
 
The following users thanked this post: salvagedcircuitry

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #58 on: September 28, 2012, 02:07:28 pm »
Thanks LaurenceW, TTi is a righteous supply!  Silly question then too, the V range is calibrated on the Maynuo, I presume?

Otherwise, I can't think of any other issue other than a bug in firmware.
Best Wishes,

 Saturation
 
The following users thanked this post: salvagedcircuitry

Offline LaurenceW

  • Frequent Contributor
  • **
  • Posts: 258
  • Country: gb
    • It's Time, Jim, but not as we know it
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #59 on: September 30, 2012, 07:08:07 pm »
So here is today's instalment of my investigation into my Maynuo M9712B electronic load, and a tiny bit of reverse engineering!

I've been impressed with the way the twin 60mm fans are driven - then only come on when they need to, and then only to the degree required. Being mid-mounted and some way from either the air intake (either side) and the exhaust (rear end) ports, they are reasonably muted, too.

A fellow poster uploaded some pictures of the M97xx series loads, so I refer you to those. Bolted to the centre of one of the heat sinks is a mystery device - an 8 Amp bridge rectifier. Now, this component alone isn't enough to manage the potential full 15 Amp load of the device, and is some way from the Maynuo's own power supply, to have been likely connected with that. What's more, I can find nothing connected to the two AC pins of the bridge rectifier, and the "+" lead is connected to ground - wtf... Very odd... Wait a minute!

Imagine a bridge rectifier symbol, but turn through 90 degrees, cathodes all pointing downwards. "-" at the top, two "AC" connections in the middle, "+" at the bottom (and connected to 0V). You now have two pairs of silicon diodes in series, all in a handy, insulated case, convenient for bolting onto a heat sink.  Now, as all electronic scholars will know (do they still teach this stuff?), a forward biased PN diode will conduct at about 0.6V, but that this drops at around 2.2mv per degree C. So two diodes in series will drop about 1.2V, reducing at about 4.4mV per degree rise in temperature.

And that's just what this rather over-rated component is doing! I can see 1.12V across it at 21C, dropping to 0.9V at 54C (which is as hot as it gets). That goes into a small circuit based on a couple of op-amps, and thence to a TIP41C TO220 NPN power transistor. This controls the current reaching the fan motors from about a 14V power supply.

And there you have it - a simple, analogue controller for the fans. They sit idle at room temperature, but just max out at full RPM after the Maynuo has been sinking 300W for six minutes. I imagine the same circuitry also sends the signal to the processor to disconnect the load, should it reach 80C, but I'm not about to try and force that...

Some more tests at FULL load, and to correct an earlier comment of mine about temperatures. Given the full 300W to sink, the MOSFETS nearest the hot end of the heat sink do get hotter than those nearest the fan, the former each reaching a case temperature of 70C after six minutes from cold, with the fans just at max rpm. Given that the MOSFET specs talk about a maximum temperature at their maximum (in this application) current of about 130C, I think they are well within thermal limits.  Indeed, the whole thermal management system seems to have no difficulty keeping pace with the dissipated heat.


Back to that TIP41C series pass element transistor which controls the fans, for a moment. The whole of the electronics are isolated from the grounded case, so in theory (and practice), neither "end" of the load + or - terminals needs to be referenced (i.e. connected) to earth. The negative lead could be at 100V with respect to earth, and the positive lead at 110V, for example. There must be a limit, here; at some point, something is going to spark over to earth. Earlier comments suggested that a true earth connection on the PCB was too close to the positive terminal (it is certainly more than close enough). But I do now also have minor concerns that the TIP41C transistor is electrically very close to the load's negative connection, but only separated from true earth (the heat sink) by a standard mica washer and nylon bush.  I have not been able to find any "official" figures for the insulation breakdown of standard T0220 insulators, but I don't think I'll be conducting any tests with the M9712B's negative lead more than 200V away from earth.

This isn't really a criticism of the Maynuo - any instrument is going to have limits to its isolation performance. But no claims either way are made for this, in the specifications - you are rather on your own.  Contrast this to my TTI power supply, which clearly states on the front panel "Outputs +/- 300V max to Gnd" So there you have it.


More on components - I notice that the capacitors in the power supply section are all G-Luxon brand. Now, I know that some readers will now be making the sign of the cross! These are not top-drawer, Panasonics or whatever. Does it matter?  But then, neither is anyone's life going to depend on this instrument running without fail for 30 years. The power supply section is linear and quite low-powered, so no nasty SMPS AC currents swishing about that might tax poorer quality caps.  At least they are 105C rated.  I am not going to lose sleep over it, and anyway they will be a piece of cake to replace, should the need ever arise.


One remaining calibration issue to sort, which is very easy to replicate and could well be a software bug. Well, this is Version 3.2, so there must have been 3.1 previous variants...  I don't think the software is going to be field-upgradable, alas :( I am in communication with Maynuo on this issue, however.

CALLING ANY MAYNUO M97xx series electronic load owners - can you get in touch? I would be interested to know if you can replicate my problem...
« Last Edit: September 30, 2012, 07:12:17 pm by LaurenceW »
If you don't measure, you don't get.
 

Offline T4P

  • Super Contributor
  • ***
  • Posts: 3697
  • Country: sg
    • T4P
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #60 on: September 30, 2012, 09:17:09 pm »
Yeah they still will always teach that stuff, but they always miss out the important details like negative temperature coefficiency.
 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #61 on: October 01, 2012, 04:43:07 pm »
Very nice further analysis LauwrenceW!  Glad to hear these updates.  For reference here's the other eevblog thread with photos:

https://www.eevblog.com/forum/product-reviews-photos-and-discussion/maynuo-m9711-programmable-dc-electronic-load-photos/msg147453/#msg147453


Maynuo's load module.

One concern is albeit the thermal management is adequate, I think it could have been better in order to reduce thermal induced accelerated aging of the load MOSFETs and increased operational reliability.  With a single fan, point A is cooled faster than B, making MOSFETs at C hotter than D; if load management was designed into the firmware or electronics, then MOSFETs at D will take up more of the work as MOSFETs at C are hotter, and have higher Rds.  Since the MOSFETs are over rated, its not a concern, but the load balance isn't equally distributed, particularly when run at its rated power.  In addition, the single thermostat at the center is a bit far from the heat sources, so the MOSFETs rely on thermal conductivity to get their heat to the sensor element, as the arrows show and there is a gap, this gap is even bigger for the MOSFETs on the other side of the heatsink.  Ideally, the fans should have been in the green circles to allow equal air flow, with added vents cut on the top or side of case.

Thermal management is better executed in the Array load modules, IMHO.  The 4 screws are the MOSFETs which are ~ equally spaced from the thermostat, and they put 2 of them instead of just one, maintaining physical equity in distances.  The module is also short, and the fan is literally in the center of rectangular frame formed by the 4 screws.  Then it has another 2 fans, one each in the rear of each module.  I've just finished a run at 130W x 4 hours, fans blowing continuously and like earlier tests, the temperature difference between the MOSFET is in 1-5oC, that is the framed silvery opening of the heatsink.

Module B tends to run 5oC warmer than A, as it picks up heat from control boards, F and G.

When the fans are on full continuously, the case interior temps are 10-15oC warmer than ambient, so far max at 40oC.  When the fans are off, the linear regulators at F & G peak at 50 and 70[ each, but when the fans kick in they drop to 40oC and 55C.


Array's load module.

I have to reconfirm on my Array, but the load modules also being a chassis ground has points of concern; its great you pointed that out.

The caps used on Array are Jamicon, not the top rated, but not notorious either.

User kmel from Germany received one not far from yours, may be good to pm him and get an update.  I'm eager to see his opinion.


Bolted to the centre of one of the heat sinks is a mystery device - an 8 Amp bridge rectifier. Now, this component ..
..controls the current reaching the fan motors from about a 14V power supply. ..to disconnect the load, should it reach 80C, ..full 300W to sink, the MOSFETS nearest the hot end of the heat sink do get hotter than those nearest the fan, the former each reaching a case temperature of 70C after six minutes from cold, with the fans just at max rpm. Given that the MOSFET specs talk about a maximum temperature at their maximum (in this application) current of about 130C, I think they are well within thermal limits. 
Back to that TIP41C series pass element transistor which controls the fans, for a TIP41C transistor is electrically very close to the load's negative connection, but only separated from true earth (the heat sink) by a standard mica washer and nylon bush.  I have not been able to find any "official" figures for the insulation breakdown of standard T0220 insulators, but I don't think I'll be conducting any tests with the M9712B's negative lead more than 200V away from earth.any instrument is going to have limits to its isolation performance. But no claims either way are made for this, in the specifications - you are rather on your own.  Contrast this to my TTI power supply, which clearly states on the front panel "Outputs +/- 300V max to Gnd" So there you have it.

More on components - I notice that the capacitors in the power supply section are all G-Luxon brand. Now, I ..At least they are 105C rated.  I am not going to lose sleep over it, and anyway they will be a piece of cake to replace, should the need ever arise.

CALLING ANY MAYNUO M97xx series electronic load owners[/i] - can you get in touch? I would be interested to know if you can replicate my problem...
« Last Edit: October 01, 2012, 05:37:59 pm by saturation »
Best Wishes,

 Saturation
 

Offline LaurenceW

  • Frequent Contributor
  • **
  • Posts: 258
  • Country: gb
    • It's Time, Jim, but not as we know it
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #62 on: October 01, 2012, 09:09:33 pm »
Hello Saturation, I can't disagree with your comments - indeed we could further point out that, in the 300W model with two heat sinks, one complete set of 4 MOSFETS are not thermally monitored at all!  One possible failure mode, therefore - the failure of the right hand fan on the unmonitored heat sink (where populated).  There is nothing (that I can see) to catch that. A second bridge rectifier bolted to the second heat sink and electrically in parallel with the first bridge would answer that (in that either hot heat sink would then invoke the 80C trip-out of the load). I may add that, in time....

Clearly, the thermal control system assumes that the power is going to be dissipated (reasonably) equally across all devices within a fair degree of margin for any error. That does so far seem to be the case.  The thermal "speed" of the system seems pretty good. I cannot see any overshoot in the case temperature of the MOSFETS, while the rest of thermal detection circuit catches up (due to thermal inertia).

I guess every design is a compromise, somewhere. Maybe in a very hot lab, on a crowded equipment shelf, running flat out for days at a time, the Maynuo might show signs of stress. Maybe. But none of those events will occur in the life of my particular M9712B!

In the Array, are the fans individually controlled by those thermostatic switches, and either on or off? At full chat, can you get it to hover?  ;D

Yes, I had seen KMEL's thread and picture posts, which is why I decided not to bother with any photos - I couldn't really better those.
If you don't measure, you don't get.
 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #63 on: October 01, 2012, 11:38:02 pm »
Hello LaurenceW,

Wowouch, you're right.  That even carries on to the BKPrecision version, unless there is another thermostat elsewhere.

Yes, I think the end purpose matters. In my need for a design the eload should run on rated load indefinitely, ultimately protected by its overheat protection circuitry [ particularly if a fan or a MOSFET should fail while on duty].  Assuming the right load module is under the same thermal constraints as the left load module without monitoring is a bit of a gamble.

Yes, on the Array, each fan has one thermostat, and its on/off.  There are 2 speeds then, call it low and high.  Since the modules are beside each other, there is some cooling cause by the thermal conduction, so the 2 modules aren't mutually exclusive in terms of the fan effects, i.e, the fans are not just additive but synergistic as each is turned on.  You can tell its fully independent because the hotter load module A tends to turn on more than the load module B, and it set for the quieter intake bottom fans first.  Hover would be nice  ;D but as bottom fans are intake, and the rear fans are exhaust, so maybe worse case the Array would fly off the desk towards me.  ;)

Hello Saturation, I can't disagree with your comments - indeed we could further point out that, in the 300W model with two heat sinks, one complete set of 4 MOSFETS are not thermally monitored at all!  One possible failure mode, therefore - the failure of the right hand fan on the unmonitored heat sink (where populated).  There is nothing (that I can see) to catch that. A second bridge rectifier bolted to the second heat sink and electrically in parallel with the first bridge would answer that (in that either hot heat sink would then invoke the 80C trip-out of the load). I may add that, in time....

Clearly, the thermal control system assumes that the power is going to be dissipated (reasonably) equally across all devices within a fair degree of margin for any error. That does so far seem to be the case.  The thermal "speed" of the system seems pretty good. I cannot see any overshoot in the case temperature of the MOSFETS, while the rest of thermal detection circuit catches up (due to thermal inertia).

I guess every design is a compromise, somewhere. Maybe in a very hot lab, on a crowded equipment shelf, running flat out for days at a time, the Maynuo might show signs of stress. Maybe. But none of those events will occur in the life of my particular M9712B!

In the Array, are the by those thermostatic switches, and either on or off? At full chat, can you ;D

Yes, I had seen KMEL's thread and picture posts, which is why I decided not to bother with any photos - I couldn't really better those.
Best Wishes,

 Saturation
 

Offline ivan747

  • Super Contributor
  • ***
  • Posts: 2045
  • Country: us
Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #64 on: October 01, 2012, 11:47:58 pm »
So pretty much every 150W/300W 360V common rectangle shape loads were traced to Xiao Tony

Go figure! ...  ;D

These are the untold stories behind this industry. I am getting a bit off topic here. I like The Amp Hour because they interview guests with stories as interesting as this.
 

Offline T4P

  • Super Contributor
  • ***
  • Posts: 3697
  • Country: sg
    • T4P
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #65 on: October 02, 2012, 08:18:35 am »
I would go for the Array but it lacks v-set  :( And i'm not sure if the 372x would live up to the 371x
 

Offline LaurenceW

  • Frequent Contributor
  • **
  • Posts: 258
  • Country: gb
    • It's Time, Jim, but not as we know it
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #66 on: October 02, 2012, 07:40:51 pm »
Saturation, if the bottom fans on your Array are intakes, that's less of a HOVER and more of a HOOVER then :)

Buuuuuuuuut seriously, I wonder if we are over-analysing the issue, here? Check back with me in ten years, and I'll let you know if the thing ever blew up...
If you don't measure, you don't get.
 

Offline robrenz

  • Super Contributor
  • ***
  • Posts: 3035
  • Country: us
  • Real Machinist, Wannabe EE
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #67 on: October 02, 2012, 08:06:10 pm »
All that so called over analysis is what educates all the EE wannabe's like myself.  Please keep it up.  :)

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #68 on: October 02, 2012, 11:42:22 pm »
A hoover? A hidden benefit, it cleans my desk while doing load testing  ;)!

IMHO, teardowns with commentary are the only reasonable methods to estimate a no-name device's quality: to live up to its spec sheet, perform safely, and guess at its working life.  Without a warranty or a reputation, we are on our own, so it would be good to know what we are up against.  Had it been a Kikusui, Agilent or Chroma eload it would be overkill, [ unless it was used and we had to do a bit of upkeep];  it would be more to admire brilliant engineering rather than looking for flaws.

There are more elegant tests available on the Maynuo that would not be easy to add to the Array, but added cooling is easier to implement.  Is the entire analysis overkill?  Not if it raises a concern in the reader to do something about it and if its all theoretical, it would still be more beneficial to an electronics forum than rants about religion and free speech going on on other topics.  If the heat discussion really has one skeptical, removing the top case and pointing a desk fan at it will substantially aid cooling to the Maynuo but its not that simple a fix to add constant voltage to the Arrays functionality.


Saturation, if the bottom fans on your Array are intakes, that's less of a HOVER and more of a HOOVER then :)
Buuuuuuuuut seriously, I wonder if we are over-analysing the issue, here? Check back with me in ten years, and I'll let you know if the thing ever blew up...
« Last Edit: October 03, 2012, 11:12:19 am by saturation »
Best Wishes,

 Saturation
 

Offline kmel

  • Contributor
  • Posts: 45
  • Country: de
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #69 on: October 06, 2012, 09:45:46 am »
User kmel from Germany received one not far from yours, may be good to pm him and get an update.  I'm eager to see his opinion.
I did Laurences test routine on my MAYNUO M9711, 150W,  firmware v3.1 .
My report resulted to following comment from Laurence: "It sounds to me as though your Maynuo load is working exactly correctly!".

--
kmel
 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #70 on: October 06, 2012, 02:02:49 pm »
Good to know kmel, and thanks for the reply.  So you do not have the CV setting error and your current range are accurate?



User kmel from Germany received one not far from yours, may be good to pm him and get an update.  I'm eager to see his opinion.
I did Laurences test routine on my MAYNUO M9711, 150W,  firmware v3.1 .
My report resulted to following comment from Laurence: "It sounds to me as though your Maynuo load is working exactly correctly!".

--
kmel
Best Wishes,

 Saturation
 

Offline kmel

  • Contributor
  • Posts: 45
  • Country: de
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #71 on: October 06, 2012, 08:32:11 pm »
So you do not have the CV setting error and your current range are accurate?
I fed the M9711 with 18V 1A and it ate 15W to regulate the voltage down to 15V if VSET=15V, even in voltage high mode when VMAX=150V. This is what Laurence asked me to do.
« Last Edit: October 08, 2012, 01:35:51 am by kmel »
 

Offline LaurenceW

  • Frequent Contributor
  • **
  • Posts: 258
  • Country: gb
    • It's Time, Jim, but not as we know it
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #72 on: October 07, 2012, 01:03:09 pm »
Yes, just as Kmel says (though I think you meant "ate 15W").

With my PSU also set to 18V 1A Max, and the load set to a CV of 15V, it pulls the power supply down to the set 15V at 1A when I select the Low Voltage range (up to 20V), but fails to draw ANY current at all when on the High Voltage range. Indeed, on the high voltage setting, I have to lower the CV setting down to 8V, before the load will begin to draw any current at all !  It is still clamping the load (to, say, 12.9V actual when 6V is commanded), and it is a real "hard clamp" too - the actual voltage across the load sticks at a solid 12.9V, for a wide range of currents.

Yet all the while, it is the "commanded" voltage which is being ignored - the actual current through the load and any voltage across it are correctly displayed. So I don't think it's a calibration issue ...
If you don't measure, you don't get.
 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #73 on: October 07, 2012, 08:03:38 pm »
Thanks, LaurenceW  & kmel, could help to email Maynuo this bug so they can get cracking at it.  With their marketing efforts fairly aggressive, my guess is they'll look into it more keenly if if 2 users complain, to suggest it can't be a one-off or user based fault.

http://www.maynuo.com/downloadfile/2012092532905455.pdf

Their latest catalog.

Best Wishes,

 Saturation
 

Offline LaurenceW

  • Frequent Contributor
  • **
  • Posts: 258
  • Country: gb
    • It's Time, Jim, but not as we know it
Re: Why is BK Precision still in business? BK8500 programmable load OEM found
« Reply #74 on: October 09, 2012, 07:28:06 pm »
OK, well a happy ending to report - for now.  Maynuo have been back in touch with me. Nowhere was it written in the (probably less than entirely clearly expressed) calibration instructions that it was necessary to calibrate the high voltage range with a >100V DC source, set to around 10mA limit current.  Bizarrely,  on Maynuo's instruction, when I recalibrated the M97 load using this source, it is now correct!  I may rewrite the calibration instructions and send them back to Maynuo.

There is still a question mark over the isolating serial>USB connector which is preventing the PC software from working correctly, so Maynuo are sending me a replacement one to see if that sorts the issue.

So my concluding comments are:
Product: Very good! Pretty damn accurate, good resolution, excellent features, clear display. The cooling system, while good and adequate, might have some vulnerabilities in the event of a fan failure, but I know how to fix it for myself, and at this price point I don't expect perfection.
Company: Very good! They could have quietly ignored me, but they did not. <thumbs up>
Documentation: er, mostly OK. Which is a shame, really, because clearly it has not been produced by a native English speaker, and even if it was, it doesn't always tell you what you need to know. This is not hard to get right, and is how far eastern companies can easily differentiate themselves from one another. And to my mind, the documentation is as much a part of the product as the box itself.

That's two-and-a-half out of three, then. Pretty good.

So to answer the original OP's headline question - Why is BK Precision still in business? Search me.


The End.

OH - PS! A couple of other things that I didn't expect my Electronic Load to be -

- A half-decent standalone DC voltmeter or amp meter, with trend logging software
- A reasonable digital resistance up to a couple of K or so
- A background room heater for those chilly UK autumn days when it's not quite cold enough to turn the house heating on (toasty, here in my "lab"!)
- A crude PWM motor speed controller (you can set on and off periods in fractions of a mS, and connect this between your load and power supply as a switch)
- A topic of conversation!
If you don't measure, you don't get.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf