Fake or Real DC to DC Solid State Relays.

[petespaco]

Several folks have been asking me the best way to temperature control the work coil of the induction heaters that I have been messing with. 
At this point, I think the best way is to use a PID controller and use the "Bang-Bang" approach of turning the DC power on and off as needed to maintain a set temperature.
  Mechanical relays can work, but they have limited life in that application and/or are pretty costly.  (We could argue that opinion somewhere else).
Anyway, I have had a hard time finding DC to DC SSR's, particularly cheap ones.
  But I did, just recently buy 2 40 amp 60 volt SSR's and 2 100 amp 60 volt SSR's .
I decided to try the 100 amp model first, just to give me some headroom, since my plan was to simply test the thing with a load going as high as about 40 amps.
  Well, I hooked it up, using a simple spst swtich to feed it a 12 volt signal and turned it on.  It worked properly with the 6 amp idle current that is normal for my 2500 watt ZVS induction heater. ---That is, when I turned the SSR "on"  the heater did oscillate normally.
Next, I added 40 amp load my putting a graphite crucible into the work coil.  Fine.  I threw the SSR's signal switch on and I got my 40 amps.  But--- when I turned the switch off, the current continued to flow.  I tried it a few times with the same result.  I powered down completely, then back up and the same result.
  I'm sure we have all heard about "fake" SSR's etc..  Right?
Well I tore this one down, thinking I simple replace its Mosfet with an IRFP260N.
No luck.
I finally did get the potting compound off of the "device", but it's labelled "JT1    A1009"  By the way, the thing is dead shorted.
The only A1009 I could find reference to is a BJT that is good for 2 amps!
Well, if that is correct. then there's no point in simply trying to replace it with a Mosfet, because I am pretty sure the gate/base driver won't work.

Moving on:
  I then installed one of the 40 amp SSR's and it has worked just fine, up to 30 amps.

My reason for posting this trial is just to see if others have more info about DC to DC SSR's.   I now have a list of about a half dozen of them in the 40 to 100 amp range, that are not horribley expensive.  So if any of you have a proven recomendation, I'd like to hear it.
(I will be putting up a video showing the ZVS heater being cycled in this way (manually) in a few days.)
There's a PID controller on the way.  I hope I picked the right one.

Pete Stanaitis
----------------

[davekni]

DC SSRs can have a tough time with anything except resistive loads.  If the load has a bypass capacitor, and the power supply does as well, then there is high inrush current at turn-on.  If the load starts with an inductor (as I'm guessing for your induction heater), then the SSR has a problem at turn-off.  It has to handle all the inductor's stored energy in avalanche breakdown.  Fortunately, this latter turn-off problem can be fixed with a diode across the load, supply-to-ground.  Any fast or schottky diode that can handle 40A peak current will work.  The diode conducts only momentarily at each turn-off event.

[T3sl4co1l]

DC SSRs use MOSFETs with photovoltaic stack optocouplers to generate the gate drive, at modest voltage (typically 6-9V) and feeble current (~uA).  The switching times are very slow: around 1ms for small devices.  Probably 10s of ms for a device that big.

40V and 40A (or more, due to inrush as mentioned) is a tremendous amount of power for a single MOSFET to handle.  Many won't handle it for this long, either due to lack of heatsinking, or higher order effects (2nd breakdown).  (And at this time scale, the heatsinking is nothing you can do externally -- it's only enough time for heat to flow from die to the plate it's bonded to.  A bigger plate gives more pulsed dissipation -- compare TO-220 to TO-247 to MAX-247 -- as does a bigger die.  Who knows what they used internally -- well, you do now, I guess!)

So a different technology is required here.  A transformer coupled gate driver would be fine.  This can be as simple as a two-transistor chopper on the primary side, and a FWB and load resistor on the secondary side.  The switch itself can be two beefy transistors back to back (tied source and gate; drains to load), rated for the load including inrush.

If there's much inrush (I don't remember offhand what, if any, bypass capacitors those modules have on them?), it may be worthwhile using a precharge resistor.  Get two SSRs and wire one in series with a power resistor.  Wire the second straight through, so it bypasses the resistor.  Sequence them so the resistor turns on first, then some milliseconds later, the main relay.  This can be done with a 555 timer and some logic.  The resistor value should be high enough that the SSR can switch it, at full supply, within ratings, but low enough that it's able to bring the load up to nearly full supply voltage.  The time delay should be a few time constants, so that the voltage has time to rise to that level.

Tim

[petespaco]

I just re read my original post.  When I said that the 40 amp unit worked okay up to 30 amps, what I meant was that it works just fine.  That one did not fail at about 3/4 of its rated load, whereas the 100 amp unit DID fail at about 40% of ITS rated load.
  That's why I was asking about the "A1009".  Now that I think about it, that A1009 is mounted on the top side of the PCboard.  Maybe I need to look underneath the board, if I can get it apart.

Pete Stanaitis
---------------

[petespaco]

I just took the back off of the 100 amp DC-DC SSR and found a C4110 BJT in there.  That is really the main "switch", not the A1009.
However, the datasheet says that C4110 is only good for 25 amps.  Why then, would they rate it for 100 amps?
I guess that qualifies it as a fake.
All three pins of the C4110 are shorted together.
 I bought 2 of these 100 amp SSR's for this test, so I guess I can hook the other one up and see if it can survive at 20 amps or so.
By the way, this circuit uses the H11D1 optocoupler.

Pete Stanaitis
---------------

[petespaco]

While preparing a video today where I am driving my 2500 watt ZVS Induction Heater with 48 volts, I cycled my 40 amp DC to DC SSR on and off repeatedly for about 15 minutes with currents between 30 amps, 20 amps, 12 amps and 6 amps with no problems at all.
It is this one:
TWTADE SSR-40 DD 40A DC 3-32V to DC 5-60V SSR Solid State Relay + Heat Sink

$12.99
Seller = TAISHUN

Pete Stanaitis
---------------

[klugesmith]

Pete, that's good news about success with SSR in your post right above here.  Before this thread I didn't even know they had SSR's for DC.  But was aware of photovoltaic optoisolators.  Got an industrial design with three of them, secondaries in series, to get an 18 volt low current supply with no switching frequency.

Might yours be a good application for an old-fashioned mercury displacement relay?
They are designed for switching 50 amp loads frequently, without contact erosion.
My small collection includes one in a box with a thyratron to control the coil;
I think it was from a garage door opener with photoelectric protection.
The one in picture has two poles and is rated 60 A, 480 V AC, 48 V DC.

[petespaco]

Re: Mercury Relay---

Maybe.
  I might still have one around here somewhere.   They were used in one of our copy machines to control a (roughly) 500 watt heater.  It had a clear glass tube to contain the Mercury, so you could see the plunger going downward, pushing the Mercury upward to "make" the contacts.  I think there was close to a pound of Mercury in one of them.

 Almost afraid to admit to HAVING any Mercury on the place.
I wonder how they are at DC arc quenching after seeing how much trouble they go to in quenching DC arcs in DC circuit breakers.

Thyratron-  there's a term I haven't used in a long time.

Pete Stanaitis
---------------. 

[petespaco]

I now have some experience using a DC to DC SSR to switch power on and off to a 2500 Watt Chinese ZVS Induction Heater.
It's all good news so far, if one chooses the device wisely.

Here are two videos that I just published:
Part 1:

/>
Part 2:

/>
They are a bit lengthy on purpose so the truly interested viewer can see exactly what's going on.  Feel free to skip around, though.

I will also be posting  these links on the ZVS Induction heater thread on this forum, since THIS thread isn't Focused on ZVS induction heating.

Pete Stanaitis
---------------

[Zipdox]

May I ask why you're recording in 480i in 2020?

[klugesmith]

What would High Definition add to Pete's presentation, except file size and transmission bandwidth usage?
Consider viewers using mobile devices with data limits. 

One thing I hate about the camera in my starter "smart phone":
have found no option to take pictures & videos with much less than maximum resolution.

As if a 4-cylinder engine would serve my driving needs, but nothing with fewer than 8 cylinders is made any more.
There's a real cost from having those un-necessary extra pistons (or megapixels).

[petespaco]

I just uploaded a short video of my teardown and analysis of the "Fake" 100 amp DC to DC SSR that has been the subject of this thread.

It is here:

/>
Pete Stanaitis
---------------