Driver Circuit for a Reciprocating Water Pump in a Medical Device

Hi Jan,

How have you been? I can't tell you how happy it makes me to see so many of the old gang still together.

Hi Gene,

Thanks for searching the switching circuits for me. I have never seen this configuration before and when ones skill level is as low as mine the first thought is that the designer had some omnicient insight that is beyond me. I find Michael's suggestion that this is a design that has evolved from its original intention fascinating. When circuits are on high quality circuit boards in very expensive equipment the thought that they may be jury rigged from previous designs didn't occur to me.

Hi Doug,

Sorry I haven't been around. I have missed you guys. It is hard to explain how life pulls us around and sends us down different paths.

When I get the boards the D26 is shorted as is the Q17. The Q15 has suffered heat damage to the extent that it may be exploded. It would be relativey easy to put a snubber diode across the load and I may try that on the next board that comes in. Can you think of any scenario where the snubber diode would cause an undesirable side effect? I will not likely get any feedback. If it solves the problem I will just never see the board again. That would be good but it wouldn't satisfy my curiousity. Thank you for your help.

If D26 is shorted and Q15 turns on, Q15 will see a large destructive current.

The snubber diodes, like D26, need to be large enough to handle repeated inductive spikes. It Q15 is driven with some sort of PWM, there will be a lot of inductive spikes, which could add up to absorption of significant energy.

I made a mistake in the connection of D1.

This is a better simulation (with the right circuit, similar diodes to yours for D1 and D2, possibly better values for the load).

The currents in D1 and D2 are identical which is why you can only see I(D1).

By adjusting the timing you can set the peak diode currents to Vsupply/Load R.

D1 and D2 are rated at 1A, Q15 is a protected MOSFET and should be fine with 4A continuous current. Q17 can handle a lot more current.

It looks as if the diodes are the weakest link but what they actually see depends on the timing, supply voltage and load parameters.

LTspice is free - I can send the model if you would like it.

MK

Thanks Michael for taking the time to further investigate. I have the LT Spice program and I would welcome your model so I can play with it myself. 

I will look into the possibility of upsizing the snubber diodes.

thanks

Hello John,

I would think that the designer was familiar with this converter topology

https://www.vishay.com/docs/91616/twoswitch.pdf

In the case of the forward converter with a transformer, it guarantees the magnetic reset of the core (with PWM less than 50%), stopping it stepping to saturation, and has a small benefit in recycling much of the magnetizing and leakage energy back to the supply (and thereby lowering the dissipation in the 'snubber' components). I suppose those might also be of benefit with PWM-ing a solenoid, but it's not an area I'm familiar with.

Are you sure there is no other connection to the top of what looks like it has to be a current-sense resistor (the 0.22R)?

LTSpice file here:

https://www.dropbox.com/scl/fo/876wu0j1vuus2e9t4yz0k/ABlNGiq6fHZOnWuOwdkZHkI?rlkey=9rs366j5r73tu3lft7f4a8rgg&dl=0

MK

Hi John,

I am glad you are back.

I hope you are doing well.