EEVblog #281 - BK Precision 8500 Electronic Load Teardown

[nitro2k01]

Hmm, 100906. Are you sure about how you read the date? Could be '10 as well. That's why the world ought to switch to ISO8601...
As for the stickers, I would guess they are put on from right to left. It would make sense to do the calibration last, don't you think? (To make sure everything is functioning AFTER the burn-in and HV testiing.)
And why didn't we get to see you turn on the unit so we could see that VFD display at least for a few seconds in the end of the video?   

[olsenn]

Awesome! I'll have to watch this video when I get home; I have the Itech 8511 version (150W) but I imagine they are near identical. I was going to take it apart but just never got around to it

[M. András]

this looks nice btw but what the hell is making the price of this so high? or just a huge markup

[olsenn]

this looks nice btw but what the hell is making the price of this so high? or just a huge markup

You can get the IT8511 for around $400 - $500 on ebay. This may sound like a lot of money to you if you are in school and don't have a job... but it is actually very cheap!

[M. András]

not in school but that 4-500 usd is my whole monthly paycheck, yeah they pay shitty in hungary, but there are hardware ans software which i dont get what the hell cost that amount of money, lots of work etc but those wont justify the 2-3x markup on things after the baseprice includes these costs

[olsenn]

not in school but that 4-500 usd is my whole monthly paycheck, yeah they pay shitty in hungary, but there are hardware ans software which i dont get what the hell cost that amount of money, lots of work etc but those wont justify the 2-3x markup on things after the baseprice includes these costs

I'm not sure the markup is quite that high actually; BK Precision needs to pay their engineers a lot more than perhaps if they were stationed in Hungary. If DC electronic loads sold like ipods they could get away with selling it for much less profit per unit, and they would get great deals from the component manufacturers for bulk orders etc, but as it stands, when you add up the price of the metal frame, the heatsyncs, PCB, components, screen, design costs, testing, marketing... something like that will cost quite a few american dollars to produce.

I'm sure you can get a Chinese e-load on ebay or something though if you're tight on budget

[Rufus]

The reason you get an isolated RS232/USB interface is that the TTL interface on the DB9 isn't isolated.

All the load internal circuitry is referenced to the -ve input. When you stick 500v on the +ve input (of the 500v version) and don't connect the -ve input all internal circuitry is sitting at 500v. Dave worries about clearance and creepage on a couple of tracks carrying mains while you should worry about clearance and creepage between every part of the internal circuitry and everything that is earthed.

Would you trust a membrane keypad with 500v on the contacts? Would you expect 500v to come out of the DB9 on the back? The remote sense terminals on the back will also have 500v. The trigger input is probably opto-isolated.

I commented in the other thread on loads about the BK manual having a page of safety drivel about mains operated equipment while not bothering to mention you can kill yourself with the DB9 on the back.

All these cheap Chinese loads must come from the same (quite old) design. My Beich also has a big bridge rectifier bolted on the heatsink. It has the two middle legs cut off and is connected with skinny wire. I assume it is used to sense heatsink temperature, hard to imagine several designers independently coming up with that solution.

[BravoV]

Just curious about those big wire shunt resistors, if they're precision and supposed to be stable (low TC), are they just bare naked thick metal wires ?

If some how they're not protected or coated, they must be rust or corrosion resistant since placed and exposed directly to the intake vent holes that humidity from the air will hit them 1st.

[free_electron]

That grind-down chip is an SST 8051 compatible machine ...
if you are going to grind it down make sure to take the logo completely off... it is unmistakable SST ( not ST !! but SST: silicon Storage Technology . These guys have been around for a long time but recently shed all their non-pure-storage devices left and right and now focus on licencing the storage technology to 3rd parties. )

And here is the kicker ... SST microcontrollers are now owned by .....wait for it ...... tadaa .... Microchip ! , yep the PIC people... i guess they finally wisened up and went for a 'modern core' like the 8051 than their 'General Instruments PIC core' ... bwahahaaaa

[Bored@Work]

I commented in the other thread on loads about the BK manual having a page of safety drivel about mains operated equipment while not bothering to mention you can kill yourself with the DB9 on the back.

Hmm, I didn't see anything like opto couplers or something else that looked like galvanic isolation in the video. Do these things really not have separated isolated power and control stages? Ugh! Bad, bad, bad.

Even if the normal ITech loads are just specified for up to 120V, which is anyhow a strange limit if you live in 240V country, I would have assumed the power stage is properly isolated.

[Rufus]

I would have assumed the power stage is properly isolated.

Which is why not documenting that it isn't is so appalling.

[robrenz]

Just curious about those big wire shunt resistors, if they're precision and supposed to be stable (low TC), are they just bare naked thick metal wires ?

If some how they're not protected or coated, they must be rust or corrosion resistant since placed and exposed directly to the intake vent holes that humidity from the air will hit them 1st.

IMO its most likely one of the nichrome based resistance alloys as opposed to manganin so it would be fairly corrosion resistant in its bare condition.   

[EEVblog]

Hmm, 100906. Are you sure about how you read the date? Could be '10 as well. That's why the world ought to switch to ISO8601...
As for the stickers, I would guess they are put on from right to left. It would make sense to do the calibration last, don't you think? (To make sure everything is functioning AFTER the burn-in and HV testiing.)
And why didn't we get to see you turn on the unit so we could see that VFD display at least for a few seconds in the end of the video?   

I always put my dates with year last. But it could be the other way around. it wasn't inconceivable that it would be an 06 design.
Yes, cal is probably the last step.
People bitch when I turn things on in teardowns.
Follow my tweets and you would have got a photo of it days ago 

Dave.

[EEVblog]

this looks nice btw but what the hell is making the price of this so high? or just a huge markup

I think $900 street price for this ($750 ebay it seems, ITech branded) is actually quite good value. They are not huge volume test gear like scope, meters, or PSU's.
The 150W version is under $500 on ebay
http://www.ebay.com.au/itm/New-1-Ch-DC-Electronic-load-0-120V-150W-BK-precision-/170639754546?pt=AU_B_I_Electrical_Test_Equipment&hash=item27baec0532

Dave.

[EEVblog]

Just curious about those big wire shunt resistors, if they're precision and supposed to be stable (low TC), are they just bare naked thick metal wires ?

Probably nichrome or some such.

Dave.

[EEVblog]

The reason you get an isolated RS232/USB interface is that the TTL interface on the DB9 isn't isolated.

All the load internal circuitry is referenced to the -ve input. When you stick 500v on the +ve input (of the 500v version) and don't connect the -ve input all internal circuitry is sitting at 500v.

I just tried it, and that is indeed correct. Not good at all...

Dave.

[tom66]

DB9 RS-232 is quite simple to isolate - why didn't they do so? Seems strange to me that they'd then supply two separately isolated cables.

[KaZjjW]

That grind-down chip is an SST 8051 compatible machine ...
if you are going to grind it down make sure to take the logo completely off... it is unmistakable SST ( not ST !! but SST: silicon Storage Technology . These guys have been around for a long time but recently shed all their non-pure-storage devices left and right and now focus on licencing the storage technology to 3rd parties. )

And here is the kicker ... SST microcontrollers are now owned by .....wait for it ...... tadaa .... Microchip ! , yep the PIC people... i guess they finally wisened up and went for a 'modern core' like the 8051 than their 'General Instruments PIC core' ... bwahahaaaa

Came to the same conclusion. Definitely SST.
Probably a chip like the SST89E564 or similar/newer. The pins 14 and 15 for the Xtal on the datasheet match the PCB. Could be the same for Vss/Vdd (pin 16/38) and Rx/Tx (pin 5/7).

[Rufus]

The reason you get an isolated RS232/USB interface is that the TTL interface on the DB9 isn't isolated.

All the load internal circuitry is referenced to the -ve input. When you stick 500v on the +ve input (of the 500v version) and don't connect the -ve input all internal circuitry is sitting at 500v.

I just tried it, and that is indeed correct. Not good at all...

My Beich has a 1M8 bleed resistor to earth, what does yours measure between the -ve input and earth?

I doubled checked the BK data sheet and manual and don't see any specification for maximum applied voltage on the inputs with respect to earth. I haven't seen that specified on any of the other Chinese clones either.

I checked the isolation on my Beich at 500v and it held up but it has no right to. In one place that isolation is provided by one thickness of pvc insulating tape separating two sheets of aluminium.

Edit: more interesting points.
All these loads are derived somehow from the same basic design, the handbook for my Beich is dated 2006 so the design is quite old.

The processor in yours does look like a 44 pin 8051 and the remains of an SST logo looks likely. My beich has an ATMega. It looks like the processor, display, and remote control protocol are the only areas all the Chinese clones significantly differ.

[EEVblog]

DB9 RS-232 is quite simple to isolate - why didn't they do so? Seems strange to me that they'd then supply two separately isolated cables.

Quite likely an afterthought.
It's safe if you use the supplied cables, but if you try and hack your own, then there could be an issue.
But of course this is no different to any similar negative terminal (mains floating) referenced instrument, but given that this is a load designed to connect to potentially high voltage sources, extra care needs to be taken.

Dave.

[nitro2k01]

What's the terminal to ground configuration? Isolated, tied one of the terminals, or connected but high impedance?
I'm thinking that maybe all the the points between the control circutry and the terminals (MOSFET base and gate, current sense, and whatever else needs to be connected) are connected with high value resistors, as to limit any stray currents going between the sections to an acceptable level. Don't know if this is a viable design. Could be vulnerable to interference, for example.

[armandas]

Follow my tweets and you would have got a photo of it days ago 

Link, or it didn't happen.

[olsenn]

Here are some quick photos I took of the IT8511 operating in CC mode (0.5A). The screen is better than these photos show just to point out. Notice that there is a low precision and a high precision mode for both voltage and current... at ~10V/0.5A all four combinations are possible, but depending on how high of a voltage/current you have, you may need to adjust the settings.

I didn't wait for the voltage level to stabalize, so the displayed values change a bit from photo to photo. Also note the way I connected the PSU through the current sense port of my DMM to the load... this may cause a slight difference in the voltage seen by the multimeter vs that seen by the load?


upload pictures


image host


hosting images


upload pictures

[amspire]

I would expect the lead and clip contact resistance to have a noticeable voltage drop. All you have to do to check is use probes to the multimeter, and make sure you measure the voltage directly to the load's terminals instead of any part of the leads or clips.

[EEVblog]

Follow my tweets and you would have got a photo of it days ago 

Link, or it didn't happen.

http://twitpic.com/9m8sto

Dave.