FIXED: Power Designs 2005A repair

[motocoder]

I have purchased a Power Designs 2005A off of eBay. While it is cosmetically in good shape, it is not working.

Upon switching the unit on, the 0-10V or 10-20V lamp (depending on the position of the switch) illuminates, but the Oven light does not come on and the meter stays at zero. The voltage across the DC+ and DC- terminals is about -0.3 volts.

Does anyone have a schematic or other technical documentation for this unit? I saw a link on one of the threads here, but the link was broken. Any advice on where to start troubleshooting?

Thanks

[robrenz]

PM me your email and I will send you a pdf of the 2005A manual

[motocoder]

PM me your email and I will send you a pdf of the 2005A manual

PM sent. Thanks!

[motocoder]

ok, internal fuse was blown. i don't have an exact replacement, but i had a smaller amperage fuse. I put that in and now I am getting something on the output. It seemed reasonably stable before i started messing with the dials, although it was about 20% below the set voltage value.

Oven light still not coming on, but that might just be the neon lamp.

I will definitely need to clean the contacts on the switches - anyone have a recommendation on how to go about that?

[motocoder]

Ok, have a few more data points now.

The oven lamp is actually fine. The SCR that switches the voltage reference heater on/off is never switching on. This is because its gate is never going high. I briefly shorted its anode to gate, and i saw the lamp light up (this might have been dumb as i may have damaged Q11)

So next I need to determine why the SCR is not being switched on. My suspicion is the thermostat is bad, or possibly Q11 (or maybe both if I damaged Q11 in my testing).

[robrenz]

I will definitely need to clean the contacts on the switches - anyone have a recommendation on how to go about that?

Link to part of the POWER DESIGNS thread here

[motocoder]

I will definitely need to clean the contacts on the switches - anyone have a recommendation on how to go about that?

Link to part of the POWER DESIGNS thread here

Thanks, I orderd some DeOxIt D100L and G100L. Where you say to first wash with a "general electronic cleaner" and to "wash the D100L off" - what do you recommend for that?

[robrenz]

I use CRC Lectra-Motive from the auto parts store.

[motocoder]

I use CRC Lectra-Motive from the auto parts store.

thanks. Also, per the thread you linked, I also need to re-lube/grease the non-contact parts of the switches where things move against each other. What do you recommend for that?

[robrenz]

Any light grease will work. I used TriFlow synthetic grease just because I had some on hand.

[motocoder]

I did some more investigation this evening. I also created a little LTSpice simulation to better understand how the heater circuit works.

Anyway, one of the things I tried tonight was to place a diode across the SCR anode/cathode, effectively removing it from the circuit. While I did this, I measured the voltage across the thermostat (terminals 9 and 11 on the the heater connection on the bottom of the PCB). The thermostat is "on" (shorted) when the meter starts up, but it never goes off. Also, the little heater tube never felt like it was getting warm at all.

So this prompted me to measure the resistance of the heater by taking a measurement across terminals 9 and 10. OPEN CIRCUIT. So I think the heater may be burned out. I also suspect the SCR is not good, but will follow up on the heater first.

BTW - does anyone know approximately what resistance I should see across the two heater terminals?

[motocoder]

Started taking the heater tube apart. Big mistake. I am wondering if the insulation inside this thing is asbestos 

[David Hess]

I would expect it to just be fiberglass.

[motocoder]

Inside the outer aluminum tube. embedded in some tape was a coil od wire, wound around the outsode of the inner tube. I wonder if this was the heater coil?

Anyone have any insight into what I should see insode this tube?

[SeanB]

That was the heater, made with nichrome wire. Inside will be the diode voltage reference. You can rewind the heater with new wire, or repair the break if you did not damage the wire undoing it. The tape is Kapton tape to provide insulation for the windings, as they typically use an uncoated wire for the heater.

[motocoder]

That was the heater, made with nichrome wire. Inside will be the diode voltage reference. You can rewind the heater with new wire, or repair the break if you did not damage the wire undoing it. The tape is Kapton tape to provide insulation for the windings, as they typically use an uncoated wire for the heater.

Ok, thanks

[motocoder]

Ok, I have ruined the heater assembly trying to get it fully apart. I thought by removing the internal screws, the tube would come off the connector, but I couldn't get it off there.  The little board that goes inside the heater tube is completely intact, but the tube will need some rework, and the heater wire needs to be replaced.

[David Hess]

I wish I was close to Seattle.

Wouldn't replacing the oven stabilized reference with a modern integrated series or shunt reference be easy and perform about as well?

[SeanB]

I have a roll of thin nichrome wire, which probably would work as a replacement heater unit if wound carefully. Want to put it in a parcel and send sea post ( not air as those are heavy) to South Africa? Sea post would probably arrive after the post office strike here is over.

[motocoder]

I have a roll of thin nichrome wire, which probably would work as a replacement heater unit if wound carefully. Want to put it in a parcel and send sea post ( not air as those are heavy) to South Africa? Sea post would probably arrive after the post office strike here is over.

I ordered some Nichrome wire from Amazon. I will give repairing it another try unless someone local here comes to claim it first.

i also need to figure out what to use to replace the various layers of different kinds of tape that were on that tube. I think the outer insulation was or could be replaced with Scotch glass cloth tape, and held together with kapton tape. The innermost layer was some sort of almost clear plasticy stuff. Probably something that conducts heat well. Silicon tape maybe?

[SeanB]

Inner was likely a PTFE clear shrink or just a sleeve, and then the wire, then kapton tape (transparent yellow) to hold it down then a covering of glass tape, then you would have a glass or rock wool insulating layer then a light binding layer of kapton to allow it to be easily inserted into the outer housing. Original might have used asbestos fibre as insulation if it was made before 1975 or so.

[motocoder]

Inner was likely a PTFE clear shrink or just a sleeve, and then the wire, then kapton tape (transparent yellow) to hold it down then a covering of glass tape, then you would have a glass or rock wool insulating layer then a light binding layer of kapton to allow it to be easily inserted into the outer housing. Original might have used asbestos fibre as insulation if it was made before 1975 or so.

Everything you describe sounds exactly like what I removed, with the exception of the innermost layer. The innermost layer was a strip about 10-15mm in width, wrapped diagonally around the tube. It was not very sticky if at all. I think the real question is a reasonably obtainable minimum set of stuff that I need to restore this thing. It will be easy to spend more than the cost of this item in nichrome wire, tape, and contact cleaner (the contact cleaner I am OK with, as I am sure I will use that on many projects)

[SeanB]

Then it was plain ptfe strip, you can use the kapton tape in it's place as it will withstand the heat of the wire.

One small roll of nichrome wire, a roll of kapton tape ( useful for many things), a roll of glass cloth tape 9 can be used to replace the glass wool if you wind a few loose layers on in place of it) and a soldering iron and solder to attach the nichrome wires to fly leads after sanding the ends with 1200 grit sandpaper to clean them. Take a small piece of sandpaper, fold in half and wipe the lead gently around 10 times to remove any oxide layer and roughen it, then solder it immediately after wrapping it around the flylead. Might not solder easily but should have good enough mechanical contact buried in solder.

[motocoder]

Then it was plain ptfe strip, you can use the kapton tape in it's place as it will withstand the heat of the wire.

One small roll of nichrome wire, a roll of kapton tape ( useful for many things), a roll of glass cloth tape 9 can be used to replace the glass wool if you wind a few loose layers on in place of it) and a soldering iron and solder to attach the nichrome wires to fly leads after sanding the ends with 1200 grit sandpaper to clean them. Take a small piece of sandpaper, fold in half and wipe the lead gently around 10 times to remove any oxide layer and roughen it, then solder it immediately after wrapping it around the flylead. Might not solder easily but should have good enough mechanical contact buried in solder.

Ok, that sounds quite reasonable and except for the glass cloth tape, is all stuff I will use again. I do have some PTFE tape here as it turns out - I use it seal threads for air and water connections. Do you think a thin layer of that would conduct heat into the tube better than the kapton?

[SeanB]

Would not trust that tape as insulation, it tears too easily. The kapton tape will do for that.

[motocoder]

Would not trust that tape as insulation, it tears too easily. The kapton tape will do for that.

Ok, I found some wrap that is used for insulating exhaust pipes. I have that and the two types of tape ordered from Amazon.com. Should be here on Wednesday.

BTW - if it turns out that the SCR is bad, do you have any suggestions on a replacement? The local electronics store has an NTE5404, which looks like it would fit the bill.
(Link to NTE 5404 datasheet: http://www.nteinc.com/specs/5400to5499/pdf/nte5400_06.pdf)

Thanks again for your help.

[SeanB]

Yes it will work, though I would use a C106D there as it will survive transients better. Just remember this heater is at mains voltage, so needs to have good insulation. Your heater resistance will have to be around 1kOhm or higher or it will heat up too fast and burn out the reference diode. Your length of resistance wire has to be long enough and thin enough so that the length wound on the inner is both covering it fully yet is insulated from shorting out either turn to turn or to the casing, and that the insulation is capable of handling 120v in operation while hot.

Myself I would modify the oven to use the 26VAC secondary and this would reduce the requirements that the heater must withstand AC mains standoff of 400V for a minute ( standard AC insulation test voltage), and then replace the pilot lamp with a LED, and replace the SCR with a C106D to handle the increased current.  Resistance will then only need to be around 60R, easy to wind with bare wire.

[robrenz]

Some info on a 2005 heater repair starting here in this thread

[motocoder]

Yes it will work, though I would use a C106D there as it will survive transients better. Just remember this heater is at mains voltage, so needs to have good insulation. Your heater resistance will have to be around 1kOhm or higher or it will heat up too fast and burn out the reference diode. Your length of resistance wire has to be long enough and thin enough so that the length wound on the inner is both covering it fully yet is insulated from shorting out either turn to turn or to the casing, and that the insulation is capable of handling 120v in operation while hot.

Myself I would modify the oven to use the 26VAC secondary and this would reduce the requirements that the heater must withstand AC mains standoff of 400V for a minute ( standard AC insulation test voltage), and then replace the pilot lamp with a LED, and replace the SCR with a C106D to handle the increased current.  Resistance will then only need to be around 60R, easy to wind with bare wire.

Ok, now this sounds like a lot more funthan winding 30 feet of fine gauge wire without having shorts between coils. I think I will look into this.

[motocoder]

Yes it will work, though I would use a C106D there as it will survive transients better. Just remember this heater is at mains voltage, so needs to have good insulation. Your heater resistance will have to be around 1kOhm or higher or it will heat up too fast and burn out the reference diode. Your length of resistance wire has to be long enough and thin enough so that the length wound on the inner is both covering it fully yet is insulated from shorting out either turn to turn or to the casing, and that the insulation is capable of handling 120v in operation while hot.

Myself I would modify the oven to use the 26VAC secondary and this would reduce the requirements that the heater must withstand AC mains standoff of 400V for a minute ( standard AC insulation test voltage), and then replace the pilot lamp with a LED, and replace the SCR with a C106D to handle the increased current.  Resistance will then only need to be around 60R, easy to wind with bare wire.

SeanB - I assume you are referring to using transformer secondary taps 3 & 4 (or 4 & 5), which give about 26VAC (RMS).

I took some resistance measurements on the old heater wire, and it matches almost exactly what robrenz posted in the other thread - 674 ohms.

Assuming my LTSpice model is vaguely close to reality, here are my calculations on heater power with the original design, and what would be required for an equivalent design using the secondary 26VAC output:

With the original design, peak heater current is 249mA, with corresponding peak instantaneous power of around 42W. Now this is a half wave rectified sine wave, and due to turn-on time of the SCR, not quite that, so if you check average power in LTSpice it is about 10 watts.

The simulation shows that the heater control circuit should still work just fine with the lower input voltage. To achieve a similar average power value with the 26VAC off transformer taps 3 & 4, we need to scale the current by 120/26, or about 1.1A, and a heater resistance of 30 ohms. Do you think this transformer can source that much current in addition to the current it is alreay sourcing for the load? Is there some way to test that? Also, would you then add a fuse to this circuit, or is the fuse on the primary sufficient?

Some info on a 2005 heater repair starting here in this thread

robrenz - Are you referring to the guy that added a microcontroller PID loop to control the temperature? Thanks, but no thanks!

[motocoder]

I think I will fuse this circuit as the consequences of the heater coils shorting could be a burned out transformer secondary or even a fire.

For the LED mod, it seems sufficient to just replace the lamp with an LED and change the series resistor to 2k2. The LED will be flickering at 60 Hz, but I don't think that will be visible.

[robrenz]

robrenz - Are you referring to the guy that added a microcontroller PID loop to control the temperature? Thanks, but no thanks!

I only meant to see pictures of the insides and the discussion on what to use for replacement insulation, not the pid mod.

[motocoder]

robrenz - Are you referring to the guy that added a microcontroller PID loop to control the temperature? Thanks, but no thanks!

I only meant to see pictures of the insides and the discussion on what to use for replacement insulation, not the pid mod.

I see. The pictures are useful. However, I didn't see where he ever posts back with resolution on insulation.

[SeanB]

The secondary winding should handle the 20W load with little issue, as it only is going to be on until hot, then it will only be on to make up losses in insulation. I would make the heater a 20W unit, so cold resistance would be around 25R, so it will take a little longer to heat up, but will get to the same set point.

You lose the standby heating capacity, but then you just leave it on when needed. Neon replaced with red ultra bright LED means you do not need to change the resistor, it will be bright enough. To reduce the loss in the winding simply connect it so the diode CR26 ( replace with a 1N4004) connects to pin 4, and the other end to pin 3 or 5. That way it will only draw current through the winding half when it is not supplying current to the main rectifier Cr13 or CR14.

[motocoder]

The secondary winding should handle the 20W load with little issue, as it only is going to be on until hot, then it will only be on to make up losses in insulation. I would make the heater a 20W unit, so cold resistance would be around 25R, so it will take a little longer to heat up, but will get to the same set point. You lose the standby heating capacity, but then you just leave it on when needed. Neon replaced with red ultra bright LED means you do not need to change the resistor, it will be bright enough.

I am not sure where you are getting 20W. Of course, my simulation results could be wrong, but based on the measurements I took on the old heating wire, and both my simulation and a back-of-the-envelope calculation, the average power of the original heater circuit is around 10W. Now peak value is 40W, so maybe that is what you are referring to? If that's the case, you are saying run the new circuit at half power? That would require a resistance of 60 ohms, not 25. 25 ohms would actually draw more power than the original.

To reduce the loss in the winding simply connect it so the diode CR26 ( replace with a 1N4004) connects to pin 4, and the other end to pin 3 or 5. That way it will only draw current through the winding half when it is not supplying current to the main rectifier Cr13 or CR14.

That is a very clever idea!

[SeanB]

20W load is around half the original heater, so will take longer to heat up but will stress transformer less from added load. Thermostat will regulate it fine. 25R cold will increase somewhat when hot, so it will start out doing a little over 30W but as it warms up it will drop likely to around 20W just as the thermostat kicks out. As the heater is only going to have half wave AC applied it will actually draw less power, probably around 15W dropping to 10W near cut out, and will hover around that with the thermostat regulating. You will need to insulate it a little better to drop the power further, but just using glass wool and an outer cover of glass wool or even just a card shield it will run with lower loss.

[motocoder]

20W load is around half the original heater, so will take longer to heat up but will stress transformer less from added load. Thermostat will regulate it fine. 25R cold will increase somewhat when hot, so it will start out doing a little over 30W but as it warms up it will drop likely to around 20W just as the thermostat kicks out. As the heater is only going to have half wave AC applied it will actually draw less power, probably around 15W dropping to 10W near cut out, and will hover around that with the thermostat regulating. You will need to insulate it a little better to drop the power further, but just using glass wool and an outer cover of glass wool or even just a card shield it will run with lower loss.

Actually, the original heater is 40W peak. If you do the math, because it is half-wave rectified average power is 0.318 of peak, or 12.7W. And in reality, because the SCR does not turn on right away, the actual average power of the original heater is somewhat less. I estimate 10-11W average power. All that is based on the cold resistance values;I assume it goes down as the heating element heats up.

So, in any case, I will stick with my numbers, which are 30R to be equivalent to original circuit, higher than that if I want to drop the peak current and power a bit.

[motocoder]

Any idea what the connector inside the heater tube is? Further inspection shows that the inside is actually badly damaged. Looks like the heater must have shorted out and generated a lot of heat when it failed. The insides are burnt and the plastic is cracking. Several connections are no longer working.

To get access, I had to destructively remove the inner heater tube. So I will need to rebuild that part. The board itself, and it's connector is fine, so if I can find the female version of this connector, I can leave the board as-is.

[motocoder]

I found an oven for another model on eBay, and got the seller to accept a reasonable offer. I'll savage the mechanical parts from that.

[SeanB]

If the donor still works sans heater will you post the remains to me, I would like to try to fix that.

[motocoder]

If the donor still works sans heater will you post the remains to me, I would like to try to fix that.

Gladly. It is the least I can do after all the help you have provided me on this thread.

[motocoder]

If the donor still works sans heater will you post the remains to me, I would like to try to fix that.

Ok, the replacement heater unit came today. I removed the board (which is very different), and verified that the connector inside is the same, and that the internal connector connects through to external pins labeled 1-8. Pin 9, 10, and 11 have a connection to the heater wires. I suspect it is the same as the 2005A unit with pin 9 and 10 going to the heater wire, and pin 11 to one side of the thermostat. The insulation on this unit was much more restrained, and I did not see anything resembling asbestos, nor any burn marks.

Question #1 : Where is the thermostat located? I am guessing it must be down inside the area with the connector. Opening that bit is how I broke the last heater assembly, so as long as its working I'm content to leave it there - just curious.
Question #2 : How would you route the connection to the NEW heater coil from the inside of the tube to the outside of the tube? The existing connections are not accessible, unless I want to try and solder to the little stub of wires from the old heater wire. I am almost certain these will eventually break, so I'd rather not try that (they also have significant resistance probably > 20 ohms combined).

[motocoder]

Disregard question #2 above; I cut away some of the plastic insulating ring at the bottom, and drilled a small hole in the PCB on the underside. With those two things I can get access to both places I need to solder the heater wires. I'll fill them both with some high temperature epoxy potting compound once it's all confirmed working.

[motocoder]

Well, heck. I put the heater wire on there (33.4 ohms). Hooked it up to a DC power supply but with no board inside. I ran it all the way up to 70C, and then measured the resistance between pins 11 and 9, but it does not change from the cold value.

So it seems the thermostat is not working in this new oven assembly. 

[motocoder]

I have a thermistor that I salvaged from a coffee maker. I think I will just make my own thermostat board out of that, a comparator/op-amp, and a 2N7000 MOSFET. Also will need a little voltage regulator and a few other components to generate a stable 5VDC for the op-amp.

Only open question is how to get the thermistor inside the tube.

[SeanB]

Look for a PTC thermal switch, you get a 70C one which is small and which will fit inside there. Otherwise you also get really small thermal switches that will fit.

[motocoder]

I'm not aware of a device like that which can be used for thermal control. I am aware of protective devices, reset-able thermal fuses, but those have a wide temperature swing and are not suitable for this purpose.

If you know of such a device, please post a link.

[SeanB]

http://uk.farnell.com/texas-instruments/lm26cim5-sha/thermostat-trip-70c-smd-sot23-5/dp/1312675

Small, but needs a 5V supply, which can easily be derived from the heater circuitry. Will need a small amount of extra componentry to drive the heater control thyristor.

http://www.farnell.com/datasheets/1771291.pdf

Page 10 of the pdf shows using it to control a heater, and you could use it to drive the thyristor in place of the transistor using a C106D, powering it off the 26VAC circuit via a single diode, a capacitor and a 5v1 0.4W zener diode and a resistor. Capacitor would be a 220uF 63v unit, zener a 5V1 400mW one and the resistor a 1k8 1W resistor.

[motocoder]

Thanks for the link. That's basically a single-chip implementation  of the circuit I was planning on building myself.

At this point I've got the thermistor all nicely wired up inside the tube. I have all the parts I need to build this circuit already here, so I'll probably just go ahead and do that rather than waiting another week for another part.

Another option I am considering is just driving a power MOSFET from the comparator directly, which means I would no longer need the SCR or transistor in the current heater circuit.

[SeanB]

That will work, and so long as the power device is rated for 50V or more it will do.

[motocoder]

That will work, and so long as the power device is rated for 50V or more it will do.

I have some IRFZ44N, N-channel Power MOSFET. They are rated at 55V.

[motocoder]

I made some progress on it this weekend, but didn't finish. Here's where things are at.

• Got the thermistor installed and measured the R vs T curve.
• Tested the heater.
• Used some potting compound to seal the holes I had to cut to solder the new heater wires to the connector.
• Created an LTSpice model for the heater control circuit (a simple comparator, thermistor, and MOSFET with some supporting circuitry like a 12V linear regulator).
• Removed the no longer needed components from the PD circuit board, and rewired things to source power from 26VAC.
• Replaced DS1 with a red ultra-bright LED and verified it is visible when powered by the rectified 26VAC signal.
• Added some standoffs to hold a little circuit board that will hold my temperature controller and the power MOSFET that switches the heater current.
• Soldered up about 75% of the little protoboard.

During this process I discovered that Q4 is bad. It was right under a hole I was drilling to mount the board, so I checked it with the diode check setting on my meter and discovered open circuit across base-emitter and base-collector. I think this transistor just drives the overcurrent lamp circuit (haven't quite puzzled out the circuit yet, but I suspect it just provides some delay to keep the lamp on long enough to see even if there is just a very brief overcurrent condition). This would explain why this lamp was not lighting when I initially tested the unit. I am planning on replacing this with a 2N5401 that I had in my parts cabinet. It looks to be a fairly close fit.

Also, I discovered that the if I run the heater coil, heating up to a thermistor setting corresponding to 50C, and then cut the power to the heater coil, there is enough thermal inertia to run the thermistor up to a resistance corresponding to around 54C. I have some hysteresis in my controller circuit, but it's sort of a moot point when the temp continues to drift up 4 degrees after power is cut. I am debating moving the thermistor off the circuit board inside the oven, and closer to the oven wall to minimize this effect. Thoughts?

This thing is definitely going to be a Frakenstein's monstor when I am done. And BTW, that Microprocessor-based PID loop makes a lot more sense to me now. It would be about the same component count as what I have now, and would do a better job controlling temps.

[motocoder]

Can one of you explain how transistor Q4, and the surrounding circuitry is working? This transistor (a 2N4888) was bad, and I've replaced it with what I think is a reasonable equivalent (a 2N5401). when I pull the current set switch out, I measure about 85v across collector/emitter of this transistor. When the current is not being limited, this transistor is in saturation and there is very little voltage across CE.

The current limiting function seems to be working, but the limit lamp is not lighting. it could be something as simple as a burnt out lamp. I'd like to understand how this part of the circuit works before i delve deeper.

[motocoder]

It's all working now!

I verified that the current limit lamp was bad. I replaced it with a yellow LED, with a 1N4001 in parallel (opposite polarity) to prevent any large reverse-biased voltage blowing the LED. The thermostat works fine, although I'm not 100% sure about the temperature. It definitely heats up a lot faster than the 10 minutes described in the manual. I'll have to take some temperature measurements later this week. If it's heating too fast, that might cause excessive temperature overshoot. I could fix that by just putting a power resistor in series with this thing to limit the heater coil current. Have to noodle on that one - might be best just to leave it be.

At 10V output, the output voltage seems to cycle over about a 100uV range, and is clearly correlated with the oven cycles. This seems to be well within specs. I did a quick zero volts and 20V calibration, and then cycled through the voltage range using the knobs. The DeOxit definitely seems to have resolved any issues there - the thing is pretty close to spot-on.

[SeanB]

Well done on giving it another half century of potential.

[motocoder]

Well done on giving it another half century of potential.

Thanks. For the stock unit, do you know what is the typical cycle time for the heater (initial warm-up, and then on/off cycle time once it's warmed up)?