Mastech Power Supply Blew Up

[MrAl]

Hi,

I rarely blow anything up these days but a quick accident led to the destruction of my $150 USD Mastech power supply.

What happened was a momentary reverse polarity connection of a 12v battery to the output terminals. I heard a loud "snap" followed by zero output voltage. Needless to say, i was very disheartened after that. This only happened to me once before when i connected an expensive meter across a very large 12v battery with the meter in the current mode setting, and that was some 10 years ago. The current function of the meter stopped working even though it could still read voltages. It is interesting however that these two blow ups have something in common which i'll get to in a second.

The next day i took apart the power supply and of course looked inside for any blown up diodes or blown up bridge transistors because that is exactly what makes that kind of 'snap' noise when it gets a huge overcurrent. To my surprise, there were none to be found! Nothing blown up, no burn marks anywhere, nothing. But what bothered me the most was on the back of the case it stated that there would be a fuse, but guess what, no fuse to be found anywhere either! I was very bothered by this because they clearly state on the case that there is a fuse, somewhere, even though they dont state where it is exactly.
After examining it for a while longer, it appears that they want the user to use a line cord (like for a PC computer power supply) with a fuse BUILT INTO IT! They expect you to use a special line cord with a fuse in it. That's totally ridiculous i think. There should be a fuse holder and fuse like everything else like that.

Ok, back to troubleshooting...
I checked the bridge diodes even though they were not burnt or broken. One read good while three looked bad. Later however i found that i had not contacted the leads of the diodes correctly from the angle i had available for the two meter leads, so i found the diodes to actually look all good.
The schematic is very complex for this thing however, and some of the parts do not show the correct value, so it became very difficult to troubleshoot with a lot of guesswork.

Finally after removing the bottom panel too, i found a very small board mounted just behind the three banana jacks used for the output. Hidden behind the board was, yes, a large diode. And yes, it was burnt and broken! Ha ha! They hid the diode behind the board so it was not very visible, AND they DID NOT show this diode on the dang schematic!! That was really a b----. If they had shown this diode on the schematic that would have been one of the first things i would have checked, because it is wired in reverse polarity orientation so that if an outside source is connected backwards the diode attempts to shunt all the current to protect the power supply.
Because they did not show that diode on the schematic and hid it under a small board that looked like it was just used for mounting of the banana jacks it cost me FOUR HOURS of troubleshooting instead of 5 minutes.

In the companies defense, the diode did protect the power supply because once it was cut out of the board the power supply worked again. But i believe this was only because the reverse polarity was only applied for a very short time. They did not use a fuse with this diode as would have been the normal way to do it, so it could have taken out the whole power supply if the diode blew open instead of short.

Now the similarity to the meter...
In the meter they did use a fuse. But it was also hidden beneath one of the PC boards so i did not find it until some years later when i went back into the meter to see if i could get the current function working again.
The meter was a different brand however.

 

[large_ghostman]

I have a MASTECH LCR meter, its fantastic, very accurate. But I got an email 2 days after buying it telling me not to use the psu supplied but battery power only! or i could damage the meter! they mentioned they would contact me when a solution was found, that was 18 months ago and never heard from them again!, I have since run the meter off a old 12V psu I had with no problems!
So dosnt look like Mastech is that great with PSU's

 

[MrAl]

Hi LG,

Thanks for the info there. I think that makes sense. I have a couple Mastech meters i like but the power supply seems a bit strange. It's not that great either because it does drift over time, something that you would not expect from a "Lab" supply.

The Mastech AC/DC current meter i have works pretty well, the only gripe i have is that it does not zero very well on the AC range. Luckily i dont need that too much anyway or else it could be a problem making lower level AC current measurements.

 

[large_ghostman]

The problem with the meter one is, they use the same psu in around 11 products!!! its rated at 2A, but shows no stability under 1.3A Load and the meter draws around 200mA. If you scope the PSU it has more noise than a school classroom at break time!!
alot of there stuff uses same parts like the cases, pSU's etc! I googled my meter and every single one had a psu issue, I prefer to use it on batteries, I have used it loads and loads but still has same bats in and still shows full power! Off hand I think it takes 8 large batteries IIRC

 

[KeepItSimpleStupid]

I have a stupid +-12-15V, 5V open frame supply that was part of something that I bought for $8.00 about 30 years ago.
I out it in a case and blew it up because of current flowing in the opposite direction,. It was the 723 regulator.

I ran into the cord fuse once. The pin of the plug twist to remove the fuse. The fuse is permanently attached to the pin, so special fuse.

 

[Reloadron]

Well MrAl on the bright side it is fixed.

Ron

 

[large_ghostman]

Was the pcb charred? if so I have done a detailed post on fixing charcoal pcb's

 

[MrAl]

Hello and thanks for the replies,

LG:
No there was no charring to be found, so that made it harder to find the part too, which was hidden to begin with.
The only sign physically was the diode itself, where half of the diode blew right off and it was bouncing around inside the cabinet when i would move the cabinet around. I was lucky i found that board, and once i did find it, i thought it might be the capacitor on the board because i still could not see the diode, and the diode was not on the schematic. It's only when i took off the board that i could see the whole top section of the board and that's when i saw the diode. I could see right away that was the culprit. I cut one lead, and then the power supply began to work again because it was just a reversed diode across the main output terminals.
How i got so lucky with this was probably because the external reverse voltage was not applied for a very long time, and so the diode blew short. If it blew open, it would have probably been worse because at least other caps would have then seen reverse voltage and current. I would have heard much louder pops and much smoke.

Ron:
Yes, well at least almost. I got the replacement diode just yesterday and plan to replace the diode, but i'll probably add a fuse too. The fuse will blow if that diode conducts ever again, and that diode can only conduct when the voltage across the output terminals is reversed which of course means an external reverse polarity source driving it (like big dumb car battery ha ha). I got a couple extra diodes and i think i might wire in more than one with the fuse too.
Too bad i forgot to order the fuse holder for this thing, i was thinking about something else and ordered a different fuse holder but it's for something else and forgot the one for the power supply.

KISS:
I cant remember if this came with a special cord, but i dont think it did. I think they expect you to use a cord based on the input you want to use either 220vac 50Hz or 120vac 60Hz. Each line voltage requires a different fuse so they try to get around that problem by forcing the user to get the cord and use a fuse that is appropriate for their line voltage. On the one hand it's a good idea, but a better idea would be to simply add a fuse holder and supply the buyer with two fuses, one for 220vac and one for 120vac. The fuses are cheap enough.

All:
The circuit for this thing is way too complex too for what it does. I think about it this way: there is ONE output, yet there is something like 200 parts. Now that's ok if it is an extremely well regulated and built power supply, but since the drift is a little too much for a power supply like this i dont think they did something right. It could even be the adjustment pots vibration because the pots they used on this thing are pure carp. A friend has the same power supply and his adjustment has gotten flaky already where you have to turn the pot back and forth several times to get it to adjust in a smooth steady fashion, otherwise the voltage jumps up and down wildly while you try to set your required output voltage...a very bad thing.


But after thinking about the fuse addition, i wonder if removing a large current from the output so suddenly as the fuse blows is a good idea. I think the power supply should be able to take that though, when the load goes from full max load to zero load in a very short time. Better than it accidentally seeing a reverse voltage i think if the diode blows open.

 

[Jaguarjoe]

Sounds like a bigger diode is in order.

A fuse will probably be too slow for you, especially a mains fuse.

You can buy IEC connectors with fuse slots, they also have cords too. IIRC, MPJA has many to pick from.

 

[MrAl]

Hi Joe,

Yes that sounds good. I was thinking of a fuse on the output where it would blow if the diode conducts. That should open the output so the reverse polarity is relieved. I hope it works ok ha ha, but that's the typical way to do it.

I was also thinking of maybe a polarity switch. That's where the connections are made with the switch in the 'off' position, and once everything looks good the switch is then turned 'on', which allows the power supply to actually power the load. With the switch 'off' it just monitors the output connection and it if is in reverse then a warning light lights up or buzzer sounds or both. Then it is just a matter of turning off the switch before making connections.

I like to use this power supply for charging lead acid batteries so i need this to be fool proof because i do this all the time.

 

[kinarfi]

Hi,
I rarely blow anything up these days but a quick accident led to the destruction of my $150 USD Mastech power supply.

What fun is that?
If you need a new design for you power supply, take a look at my new design for my variable power supply, It loves to eat TIP 35 & 36 transistors, but my massive 163H72 trasistor just keep hanging in there.
Kinarfi

 

[HiTech]

You can buy a used HP, Power Designs, or Sorensen power supply for less that half of $150 and they will operate faithfully and are bullet-proof to no end. Most of the Asian produced test equipment is feeble.

 

[KeepItSimpleStupid]

Come on Mr. Al. You don't want the transistor to blow to protect the fuse just like the old axiom.

Or
Keyur's second law:
Protecting device will work in such way that it can have maximum damage to the original circuitry.

Which is a member of the Murphy's Laws for Electronics here:

https://keyuracharya.blogspot.com/2011/07/murphys-laws-for-electronics.html

 

[MrAl]

Hi,

Well i finally got the fuse and fuse holder installed.

KeepIt:
Well, that's not always true and i found out that a circuit without a fuse always blows up the transistors but a circuit with a fuse only sometimes blows out the transistor. It's an old rule i think came about when fuses were slower and also before people knew how to use them properly.

For example, we cant just install a fuse anywhere and expect it to protect everything. It has to be done very judiciously where we incorporate the fuse into the design itself, knowing it's limitations and the limitations of the parts. Some parts can take a beating for a short time, while others can not. But the main thing is if we use a fuse it has to be treated as part of the entire design, not just as an added feature.

Lucky for my power supply though it had a reverse protection mechanism already, even though it wasnt ideal. So all i had to do in this case was add a fuse, but let me explain this in more detail because it's not apparent.

The diode is connected in anti parallel across the output terminals of the power supply. So if the user accidentally connects another source (like a battery) across the output in reverse (wrong polarity) then the diode conducts heavily and blows short. That's the idea behind the protection diode. What they did not consider though was what happens if the diode blows open. If the diode blows open it no longer protects anything because it's effectively not there anymore; not part of the circuit anymore.

So the solution then is to do what is usually done for this kind of thing for circuits that are not exceptionally high current (this one is only 10 amps). That is to place a fuse in series with the diode, so that if the diode conducts the fuse blows and the diode and circuit is disconnected from the output so the current stops.

But there is another reason for doing this that has nothing to do with actually protecting the power supply from blowing up. It has more to do with limiting the *severity* of the explosion. As i am sure you know, the longer we apply power to a circuit the more energy it absorbs. The less time, the less energy. So by limiting the time the fault can exist we effectively limit the damage to the circuit, at least to some extent. For example, sometimes the transistors blow but at least the circuit board traces do not blow off the board (vaporize and disappear). So we might end up with something we can at least repair later on even though we still have to replace the transistors. So that is part of the action of the fuse: to limit the fault current flow time period.

I have seen many transistors saved however, and some of them were 100 dollars or more each. This however was done with an underrated fuse so that the transistor could get a large current which would quickly blow out the fuse. For example, a 1 amp fuse and 10 amp transistor. Fusing time is highly dependent on the current level. The time does down fast as the current goes up. So a fuse that will blow at maybe 1.5 amps will blow much quicker at 10 or 20 amps. This could easily save a transistor duing the first run up of a power converter for example.

In my case the diode blew short, so that protected the transistors. With the fuse the outside circuit will be disconnected and therefore limit the time the current has to blow stuff up or burn stuff up. This does not guarantee that the transistors get protected (as the diode alone did not) but it does add protection against more damage being done.

Back when i was working in the field this was a more important topic because back then higher current transistors were 100 or 200 or even 300 dollars each (USD).

 

[KeepItSimpleStupid]

I took apart an APC surge suppressor that I got for free, so no harm in figuring out what;s wrong, right. Outlets worked, but the "something's wrong" light was on. Traced that problem to a thermal fuse. Jumping it restored operation. Got the exact replacement from ebay. It's waiting to be installed. Strip is taken apart in a box. So, the thermal apparently "protects" the MOV's. None are shorted or show signs of discoloring.

Since I don't know the history, the main fuse on the strip could have opened.

---

At work, I must have saved up some money, buy adding "3 AG load fuses" in series with 10 to 25 AMP Semiconductor fuses which were like $30 a pop AND hooking up the current limit control. Shorts were inevitable, since the Phase Angle Fired SCR unit operated a Variac and the wiring was exposed in a vacuum system. Later, I managed to re-design the entire system to use DC power supplies connected to programmable temperature controllers. I stuck my neck out on that one. It was just hard finding power for Seven 1500 W power supplies.

Newer models were 1 RU and operated from 95-250 V. Seven of the original supplies were 4 RU and 1/2 rack, so you could either use 4 or 2 in the same space.

It would have been nice if the newer supplies had a power readout.

 

[Zerotolerance]

This is why I only buy Fluke products

 

[MrAl]

Hi,

Ha ha, well they can be quite expensive unfortunately. That's why not everyone buys them. Nice to have though.