Dead output transformer

[cdemike]

First post from a long time lurker. I finished wiring up a 50 watt Hiwatt SAP clone using a Mojotone DR504 kit as my base. I followed the startup procedure that Rob Robinette outlined using a lightbulb current limiter but not a variac. Thankfully I discovered some issues with the bias circuit thanks to the current limiter, and I was able to bias the output section, though the EL34's that came with the kit were pretty seriously mismatched and the B+ was very high at 520V. I was able to bias one one at 20ma and the other at 35ma and figured the B+ might "settle in" when running (I've gotten this recommendation in the past, but after this experience I'm thinking that might be bad advice). It worked and sounded pretty good for about 5 minutes, but it suddenly started blowing the 1A power transformer center tap fuse. After troubleshooting, I've discovered that my output transformer center tap has continuity to ground and to the plates of either output tube socket. There's also continuity between either socket's plates. I know continuity to ground is unusual (and dangerous!), but I checked my wiring and the power section seems right based on the layouts from Mojo, Mark Huss, and the schematic. I ran the amp every time with a load attached (a Weber MASS 100 set up as a dummy load, i.e., with all the knobs at 0) and with the impedance selector always matching my load. When I played the amp, I used the attenuator, and again kept it at 8 ohms.

So I'm trying to figure out what went wrong here before I either throw out a perfectly good output transformer because I made a stupid wiring mistake or roast another output transformer. I've attached some pictures and my schematic and am totally happy to upload more pictures (pardon the sloppy stray wires in the post-disassembly pictures!).

Edit: forgot to attach the schematic.

[martin manning]

When you found continuity to ground from the OT CT, did you have the CT disconnected from the standby switch? Did you check the bias supply voltage at the power tube grid pins (pin 5) to insure that it has enough range? Does it measure the same on both power tubes?

[xtian]

Welcome, Mike! First, note that the OT primary (brown/red/yellow wires according to your schematic) is a continuous copper wire with a center tap. So, measuring plate to plate on your power tube sockets should show maybe 40 ohms resistance, and maybe 100-200K ohms from plate to ground. If you disconnect and isolate the brown, red and yellow primaries from everything, there should be no continuity between any of these wires and the OT's iron laminations, or to any of the primary wires. Can you confirm this stuff?

[Stevem]

Unbolt the OT from the chassis and shim it up off the chassis with some card board, then make sure your standby switch is open and then check to see if you still have a dead short to ground, if so make the same test with the CT unhooked from the standby switch .

I have seen a lot of shorted OTs in my time, but never one that was shorted to its iron core.

[R.G.]

+1 to stevem's comment. I've seen exactly one transformer shorted to the core, and it was on a large computer's power transformer, not in a guitar amp.
also,
+1 to xtian's comment. Continuity is not necessarily a short. Follow his advice about isolating the primary and measuring for continuity to ground with the CT disconnected.

Shorted transformers do happen, but I think it's more likely that you have a shorted rectifier or filter capacitor. A shorted filter cap will make the CT show continuity to ground and cause fuse popping and other maladies, as will a shorted rectifier.

[cdemike]

I'm blown away by all the quick help from this community. Thank y'all so much -- glad I joined! I am embarrassed to admit I had a no connection between the OT secondary CT and ground (using the Mojotone layout [https://www.mojotone.com/Amp_Kits/Briti ... _50_WD.pdf], I omitted the black wire to ground from the left output jack's sleeve). I I did all my tests with NFB set up and connected, so there might be a clue.

When you found continuity to ground from the OT CT, did you have the CT disconnected from the standby switch? Did you check the bias supply voltage at the power tube grid pins (pin 5) to insure that it has enough range? Does it measure the same on both power tubes?

Sorry I should have mentioned the setup for the measurements. When I measured, the output section was mostly taken apart: no tubes in, with the OT primary CT unsoldered and measuring the CT wire on its own. So isolated OT CT to isolated primaries (assuming no issues with the sockets). The bias pot had good range at -35 to -72VDC when running with no tubes in and was equal on both grids.

Welcome, Mike! First, note that the OT primary (brown/red/yellow wires according to your schematic) is a continuous copper wire with a center tap. So, measuring plate to plate on your power tube sockets should show maybe 40 ohms resistance, and maybe 100-200K ohms from plate to ground. If you disconnect and isolate the brown, red and yellow primaries from everything, there should be no continuity between any of these wires and the OT's iron laminations, or to any of the primary wires. Can you confirm this stuff?

Thank you so much! Just got some measurements done and found 45.7 ohms between plates. There is no direct connection between the primary or secondary sides with or without the chassis bolted to the OT. There's also no connection between the secondaries and ground (thanks to my bad joint). I did see what happened if I had a good connection between the speaker jack sleeve and ground and found continuity between all secondaries to ground with a cable plugged in, OT unbolted, and OT insulated from the chassis. There is no continuity between the secondary taps without the jumper between the secondary CT and ground.

Unbolt the OT from the chassis and shim it up off the chassis with some card board, then make sure your standby switch is open and then check to see if you still have a dead short to ground, if so make the same test with the CT unhooked from the standby switch .

I have seen a lot of shorted OTs in my time, but never one that was shorted to its iron core.

Just followed this and found that the resistance from the isolated OT primary CT to ground is 350 ohms. When unbolted, there's no connection between the primary CT to ground or to the secondaries, with or without the corrected jack sleeve connection to ground. I measured resistance the primary CT at different points on the OT and found that on the side closer to the middle of the amp the resistance gets as low as 80 ohms. When I traced my probe from the outside-facing side of the OT toward the inner side, the resistance fell. There's no continuity between the secondaries and the housing.

+1 to stevem's comment. I've seen exactly one transformer shorted to the core, and it was on a large computer's power transformer, not in a guitar amp.
also,
+1 to xtian's comment. Continuity is not necessarily a short. Follow his advice about isolating the primary and measuring for continuity to ground with the CT disconnected.

Shorted transformers do happen, but I think it's more likely that you have a shorted rectifier or filter capacitor. A shorted filter cap will make the CT show continuity to ground and cause fuse popping and other maladies, as will a shorted rectifier.

Given how rare a short between the primary winding and the core, I'm hesitant to pronounce the OT dead, but I'm not finding continuity to ground at any of the capacitor leads and ground. I'm struggling to think of an explanation of how my CT has continuity with the OT housing and laminations, though. Thank y'all again for your patience walking me through this.

Edit: clarity

[R.G.]

It might be time for a more complete test on the OT. To do this, Disconnect all of the OT primary and secondary leads, doing any marking you will need to put them back. This may be redundant, given your earlier testing, but for solving mysteries, sometimes it's best to start from absolute zero, eliminating merely improbable things for certain.

With all the leads floating in the air, check the resistance (not continuity) from primary winding to secondary winding, expecting it to still match your earlier measurement. Yeah, this is redundant, but it reaffirms what you already did, and it's easy.
Test primary and secondary to the mounting bolts on the chassis. This might be redundant, but having all the leads disconnected ensures that there's no sneaky path you forgot, like I have forgotten before. Both measurements should be over range on the biggest resistance range of your meter. If the primary still shows continuity to the transformer mounting on the chassis, it has some internal problem and should be replaced. With all leads in mid-air, there should be no continuity to the shell or mounting at all.
Sometimes there are manufacturing defects. The one short-to-laminations I mentioned was such a fault. The wires on one of the layers had not been fully impregnated with varnish and vibration had worked one wire loose on the edge. It migrated out to the edge of the layer insulation and touched the laminations, and vibration proceeded to wear through the wire insulation.
Shorts to lamination are funny. The black coating on the laminations is one of the iron oxides, deliberately grown there as an insulator between laminations. It's an unspecified resistance, and is there just to keep core losses down. The resistance from lamination to lamination and to the shells and mounting is not easily predictable, but could be in the range of the 350 ohms you've mentioned.
But all-leads-in-mid-air removes the issue. If the windings are not very, very high resistance to mounting, even if it's a couple of K, the transformer is defective. Leads-in-air also eliminates other issues like bad joints.

[cdemike]

It might be time for a more complete test on the OT. To do this, Disconnect all of the OT primary and secondary leads, doing any marking you will need to put them back. This may be redundant, given your earlier testing, but for solving mysteries, sometimes it's best to start from absolute zero, eliminating merely improbable things for certain.

With all the leads floating in the air, check the resistance (not continuity) from primary winding to secondary winding, expecting it to still match your earlier measurement. Yeah, this is redundant, but it reaffirms what you already did, and it's easy.
Test primary and secondary to the mounting bolts on the chassis. This might be redundant, but having all the leads disconnected ensures that there's no sneaky path you forgot, like I have forgotten before. Both measurements should be over range on the biggest resistance range of your meter. If the primary still shows continuity to the transformer mounting on the chassis, it has some internal problem and should be replaced. With all leads in mid-air, there should be no continuity to the shell or mounting at all.
Sometimes there are manufacturing defects. The one short-to-laminations I mentioned was such a fault. The wires on one of the layers had not been fully impregnated with varnish and vibration had worked one wire loose on the edge. It migrated out to the edge of the layer insulation and touched the laminations, and vibration proceeded to wear through the wire insulation.
Shorts to lamination are funny. The black coating on the laminations is one of the iron oxides, deliberately grown there as an insulator between laminations. It's an unspecified resistance, and is there just to keep core losses down. The resistance from lamination to lamination and to the shells and mounting is not easily predictable, but could be in the range of the 350 ohms you've mentioned.
But all-leads-in-mid-air removes the issue. If the windings are not very, very high resistance to mounting, even if it's a couple of K, the transformer is defective. Leads-in-air also eliminates other issues like bad joints.

I ended up deciding to completely pull the output transformer to make absolutely sure before a big purchase like an output transformer. Here's what I found:

• All primary taps have continuity to the frame where the mounting bolts attach the OT to the chassis, but only on the primary side. The brown tap and center tap (red) have 26 ohms each to the frame, and the yellow tap is 46 ohms to the frame.

• There is no continuity between the primary and secondary taps

• There is no continuity between the secondary taps and the frame

• The impedance between the primary taps seems normal with the center tap reading 20 and 26 ohms to the brown and yellow taps, respectively. There is 45 ohms resistance between the brown and yellow taps

• The secondary center tap (green) reads 1 ohm to each of the other secondary taps

• Excluding the secondary center tap, each of the other secondary taps has less than one ohm resistance between them (purple to orange, purple to white, orange to white)

I've read about floating ground on output jacks and know it's a no-no with any kind of NFB set up, but I'm pretty surprised that omitting the ground reference seems to have totally roasted my primary. In any event, I'm never forgetting to ground my output jacks on an amp with NFB again. To confirm this does mean my OT is dead right? If so, I'll order a new one, but is there anything else that may have died alongside the OT? I'll start going through anything else that may have been collaterally impacted, but I'd really appreciate any pointers where to pay special attention.

Edit: Here's the Mojotone output transformer diagram:

[Stevem]

You should not have any resistance from the primary winding to the structure of the OT.

This means either the Trans is shorted and now toast, or maybe you got lucky and since this repro OT has bells on it then maybe one of the primary leads has got cut and is shorting too the bell.

It’s easy to remove the bell and have close look!

[martin manning]

I ended up deciding to completely pull the output transformer to make absolutely sure before a big purchase like an output transformer. Here's what I found:

• All primary taps have continuity to the frame where the mounting bolts attach the OT to the chassis, but only on the primary side. The brown tap and center tap (red) have 26 ohms each to the frame, and the yellow tap is 46 ohms to the frame.

• There is no continuity between the primary and secondary taps

• There is no continuity between the secondary taps and the frame

• The impedance between the primary taps seems normal with the center tap reading 20 and 26 ohms to the brown and yellow taps, respectively. There is 45 ohms resistance between the brown and yellow taps

So the primary seems intact, but shorted to the frame somewhere. I would have thought the brown would have zero ohms to the frame given that you see 46 end-to end, 26 from red to the frame, and 46 yellow to the frame.

• The secondary center tap (green) reads 1 ohm to each of the other secondary taps

• Excluding the secondary center tap, each of the other secondary taps has less than one ohm resistance between them (purple to orange, purple to white, orange to white)

I've read about floating ground on output jacks and know it's a no-no with any kind of NFB set up, but I'm pretty surprised that omitting the ground reference seems to have totally roasted my primary. In any event, I'm never forgetting to ground my output jacks on an amp with NFB again. To confirm this does mean my OT is dead right? If so, I'll order a new one, but is there anything else that may have died alongside the OT?

I don't see how leaving the secondary ungrounded has anything to do with the primary short. I suspect it was there out of the box, or if you disassembled it you may have created the short when reassembling it.

[Stevem]

What’s odd here is that the insulation used to keep either primary or secondary winding from shorting to the core ( if the trans is made right) is far better then the baked on enamel coating on the winding wires themselves.

Like I posted, I would pull that end bell and investigate.

[cdemike]

Just pulled the end bell and was greeted with the smell of burnt enamel and evidence of an arc where the brown lead connects to the primary winding on the bell (picture at the bottom of this post). I saw there was a corresponding burnt spot on the paper wrapping over the primary leads, so I decided to investigate.

Before I get into what I found while tearing down what may have been a repairable transformer, I wanted to say I reached out to Mojotone when things started pointing to a bad OT, and that they offered me a discounted new output transformer. I decided to run with that since I wanted to pick up some small parts to clean up some sloppy layout choices on my end. I also wanted to get a 3-position on-off-on standby switch along with a choke so that the amp could be switched to use the usual Hiwatt-style 100R resistor right after the standby switch or a choke like the Jimmy Page SAP amp. Mojo's shipping fee is $30 for orders under $100, and those things together would have been about $60. I had the choice to order those things and pay shipping from AmplifiedParts or similar for stuff I wanted to get anyway, or save on shipping while also simplifying getting those parts all at once. Plus, I liked the added peace of mind that I'll be reasonably sure that there's no B+ short to the chassis when I'm not really confident yet about what caused the OT to fail.

TLDR; ordering a new output transformer that helped me avoid the big shipping fee while also being more confident that I and the other parts in the amp are safe(r) than with a shorted OT primary. So I treated what follows as a learning experience to dig into how the OT is put together. Stevem was right: the brown tap's connection to the primary winding had gotten hot and shorted to the bell. This was the first time I took this or any transformer apart, so thankfully I can spare myself the frustration of knowing this short was a result of a bad OT re-assembly on my end. I peeled the paper covering off and confirmed the dark spot I found was indeed a hole allowing the primary to short to the bell and then to ground (also pictured at the bottom).

So this gives a clue as to what happened and I had the opportunity to learn more about how transformers are put together and how to take one apart.

I don't see how leaving the secondary ungrounded has anything to do with the primary short. I suspect it was there out of the box, or if you disassembled it you may have created the short when reassembling it.

When I reached out to Mojo, they said running the amp when the secondary doesn't have a path to ground is the same as running the amp without a load. I don't really follow the explanation. My electronics background is from experience working on guitar equipment and from some physics classes in college, so I'm not trying to say I disagree, but rather that I'm not qualified to weigh in confidently. By the same token, I'm definitely not trying to cast doubt on your post, Martin. Does this explanation make sense, or is there some other cause that jumps out to y'all given this type of failure?

[martin manning]

When I reached out to Mojo, they running the amp when the secondary doesn't have a path to ground is the same as running the amp without a load.

That is true if the ground is necessary to connect the load across the secondary. A missing secondary ground reference will just result in some excess noise.

[cdemike]

When I reached out to Mojo, they running the amp when the secondary doesn't have a path to ground is the same as running the amp without a load.

That is true if the ground is necessary to connect the load across the secondary. A missing secondary ground reference will just result in some excess noise.

That makes sense to me, since the effective load is still the speaker. So without a reference to ground, the potentials should be the same across the OT secondary since in any event the load is in series, and adding a parallel path to ground should just act as a reference like you said. What I'm struggling with is there seems to be either a problem from the very beginning or something caused the hot spot in the paper where the tap-winding connection burned through to the bell. Since the amp was okay for a few minutes after first running, I'm trying to think of what could have caused the point in question to get hot enough burn through to the bell. If it was a defect, perhaps it was just thin to begin with and the regular primary operating voltage caused the the insulation failure?

Martin, it sounds like you still think it may have been like this from the start? Or is there a potential culprit in my amp that could have caused a surge and a subsequent insulation failure? Since you and Stevem predicted this type of short, I looked up what may have happened to others and came across this: https://ampgarage.com/forum/viewtopic.p ... 56&p=47720

So it clearly happens because of structural issues, but is there anything I may have to investigate before putting in a new OT?

By the way, I really appreciate all the help so far. I definitely regret not joining sooner! Looking forward to contributing to the community how I can.

[pdf64]

Without a 0V common reference for the OT secondary, the potential divider of the NFB network may not have had the expected effect. Too much NFB can result in the feedback actually being positive, sending the power amp into oscillation, possibly ultrasonic.
That could overheat the output valves, screen grid resistors, and the OT.