BK precision 1760A design flow?

[isleder]

This is my first post here. I hope it ends up on the right forum.

I recently acquired on ebay a BK precision 1760A triple output power supply. It was sold as is for repair. When I turned on the unit, the 30V outputs displayed and measured 44V. The 4-6.5V output works. My initially thoughts were that the power transistors were blown.
When I opened the case I soon realized that it is an old school design: LM723, LM324, 2N3055. I have designed power supply like this in the 80s. I did not know that after 30 years these chips are still in use in commercial offerings!

After a visual inspection I noticed that 2 of the LM324 chips were blown: one literally the other melted the dip socket. The 2N3055 transistors tested out fine. I measured 38.5V between pin4 and pin11 on both LM324. I thought this can't be right, because these chips are rated for 32V.
I started investigating wether there is a failed voltage regulator for these supplies. I did not have a schematic diagram and had to take the board out tracking the traces. For my surprise there are no regulators for the + and - supplies for these chips, but a simple split transformer design with a bridge rectifier and caps. I measured 2 x 14.5V AC on the transformer outputs and 2 x 19.5V DC on the caps. My AC mains is 124.5V, which is more than what the supply is rated for 115V, but I expect a power supply to be able to handle a bit of over and under voltages.

I am not sure how the LM324 chips supposed to survive those voltages and how did they design them with such a high voltage rails?
Even if I derate the AC mains to 115V still the chips would be over powered.
My question is wether there are high voltage LM324 chips, or the transformer windings are just wrong or is it a design flow? I was thinking of adding +-15V regulators. But I am not sure yet how much the LM324 chips need to swing on the outputs.  I could not find the schematics or the service manual for the power supply online.

Thanks
Istvan

[SeanB]

Most opamps like 741 and 324 are designed for a 30V rail, and have an absolute max of typically 36V. However certain manufacturers have made them with a 44V abs max range, so the will likely work there. Just change them with new ones, and if you are worried get ones that are MIL spec rated in ceramic cases ( likely now 30 year old NOS units. Those will have been burnt in at 44V, so will work well there) and use a turned pin socket to mount them in, not a stamped cheap socket. I have used bog standard 741's in a TO100 case with a small push on heatsink to replace OP27's that were running from a 22v split supply, as they had to give an output that swung at least 16V above ground either way to drive the other end. They ran slightly warm ( warmer than my finger) without the heatsinks on them but as I had them to hand and the chips ran at only slightly above ambient they stayed on.

Likely the designers relied on the chips being able to operate beyond the abs max, as this is a limit where they will have a certain number of failures but typically a good portion of then will be functional beyond to some degree, and you got some that ran for years before failing. Just replace them and see if there is any other damage and if not keep some as spares just in case.

[isleder]

Thanks for the reply. I ordered a bunch of LM324s.
It is an interesting idea to run the chips in over voltage mode on purpose.
 
I did a bit of reverse engineering last night and the ground of the LM324 supply is connected to the power supply positive output. The series transistor 2N3055 is driven by TIP41C in a darlington configuration and that is driven by a small NPN transistor with a pull up resistor to the 19.5V. If the LM324 fails then the output transistor is driven high to the 44V rail by this pull up resistor. If I just simply replace the LM324s then it would still be a time bomb, not knowing when the chip fails and drive the output to full rail voltage.

The LM324 output does not need to swing all the way to +19.5V, because it just need to be able to drive the 3 NPN transistors through the diodes, maybe up to 3V max. My plan is to add 15V regulators and replace the chips afterward and see how it works.

[isleder]

I replaced the two LM324 and added voltage regulators for the +- rails, just to be safe (no overheating).
All is well and now it works fine.

[pmcouto]

Hi isleder,

I’m glad your PSU is now working fine.

It seems that RIGOL is not alone designing “time-bomb” PSUs…

[isleder]

I still cannot understand why it was designed this way. It is really a basic idea to add safety margins to a design. Adding a couple of basic regulators would not have increased the price that much. The LM324 does not require high voltage rails in this case, even +- 6V rail would have sufficed.

I did not mention but the 7 segment display regulator heat sinks are quite hot too (cannot touch it by hand). I left those alone for now, as it does not affect the supply operation if they blow up.