Magnavox 9302 - Rebuild

[Cameron06]

Regarding the 12AX7s, I've tried both with them installed and removed. It doesn't make a difference.

Here's a link to a few photos of the underside of the amp; I added a photo of the top, too, to help with orientation since I moved a few things around. I tried multiple times to post the photos here, but it keeps saying they're too large even after I scale them down, so I dropped them in my Google Drive instead.
https://drive.google.com/drive/folders/1pQC1qMGlqeOZxxp9S8INFrZw0H8NNwWo?usp=sharing

I also exported all my schematics and included them here as well.

I may try just tracing each connection on the amp and marking it off on the schematic as I find it, but I've already gone over the wiring so many times, this feels kind of futile.

Thanks again for the help; if you weren't on the other side of the country, I'd buy you a beer.

(I'm aware how messy the underside of the amp is, btw. I plan on cleaning it up once I actually get it running.)

Edit:

I just noticed I mislabeled "EFB" as "AFB" in one spot on the schematic. The two items are the same, though.

 

[gadget73]

I really don't like how long the leads are on those coupling caps, or that they lay across the wiring to the output transformers. I'd much rather see those short and as direct from the inverter tube to the output tube as possible. Where they are could cause oscillation problems. Same with the caps between the buffer and drivers. Pic of mine attached for reference, leads on everything is as short as I could get by with.


problem with those Dale resistors is they're impossible to get values from unless they are aimed exactly the right way. I honestly have no idea if things are in the right spot so I'll have to take your word on that.

something doesn't look right in the power supply. I can't quite tell but it looks like you have the rectifier running to the choke without also being connected to a cap right off the rectifier. That will make the voltages low in the supply which might explain the odd EFB voltages. It will also make the choke extremely unhappy.

on the schematic you have the blue and brown output transformer leads flipped. The cap that ties into the feedback comes from the brown one, which is in-phase with the 16 ohm tap. Reversing those will make it scream bloody murder from inverted feedback.

 

[Cameron06]

@gadget73

Thanks for taking a look.

1) Regarding the capacitor leads, I hadn't even considered oscillation issues. Looking at your amp, I can definitely see how they're way too long. It also adds to the general clutter, too. Once I have some time off work, I think I'm going to disassemble the amp and rebuild it with a more logical layout. My problem is that I installed the resistors first, then the caps, then the transformers, and it really ended up being kind of a situation of "install stuff wherever it fits."

2) I'm very certain that the resistors are the correct value, but yeah, the Dale resistors don't have stripes, so they're impossible to read. I went with those because they were the best I could find on Mouser that had a low tolerance and a 500V working voltage.

3) The rectifier runs to a 50uF capacitor and to the choke. The clutter in the amp probably makes it really hard to see. I attached a picture with this highlighted.

4) It's interesting you bring up the Blue/Brown plate connections on the OPT. Originally, I had these reversed, then I swapped them, although I had them this way for a little while, which may have caused some of the initial problems. The brown lead *does* lead to the feedback now, although my schematic still shows them reversed. Unfortunately, it's also very hard to see this; I attached another picture with this highlighted. I have a sneaking suspicion that having these reversed for a while may have caused some damage to the tubes. One of the tubes is melted for sure; it's internally shorted.

So, my plan at this point is:

• Tear out everything inside the amp. I have spares for almost everything except the coupling capacitors (because these are so expensive), so most components will be fresh.
• Rebuild the amp with all leads as short as possible and eliminate and crossing of leads.
• Re-check EFB voltage using a 15k resistor to ground.
• Assuming the EFB voltage is correct, install a new set of EL84s and try again.

It'll be a few days because of my work schedule, but I'm looking forward to tearing this apart and having a fresh start. I've done a lot of electronics building in the past, but for the most part, it's been almost entirely digital. One thing this project has really showed me is that analog is a lot of fun. There are a whole new set of challenges when you're concerned with more than just making sure 1s and 0s get from point A to point B. Once this amp is finally up and running, I think I might try putting together a phono preamp next.

 

[markmcmuz]

I just ordered a pair of these edcor transfos for my second 9300 DG rebuild. I gave up on waiting for the dynakit ones.

 

[markmcmuz]

@gadget73

Thanks for taking a look.

1) Regarding the capacitor leads, I hadn't even considered oscillation issues. Looking at your amp, I can definitely see how they're way too long. It also adds to the general clutter, too. Once I have some time off work, I think I'm going to disassemble the amp and rebuild it with a more logical layout. My problem is that I installed the resistors first, then the caps, then the transformers, and it really ended up being kind of a situation of "install stuff wherever it fits."

2) I'm very certain that the resistors are the correct value, but yeah, the Dale resistors don't have stripes, so they're impossible to read. I went with those because they were the best I could find on Mouser that had a low tolerance and a 500V working voltage.

3) The rectifier runs to a 50uF capacitor and to the choke. The clutter in the amp probably makes it really hard to see. I attached a picture with this highlighted.

4) It's interesting you bring up the Blue/Brown plate connections on the OPT. Originally, I had these reversed, then I swapped them, although I had them this way for a little while, which may have caused some of the initial problems. The brown lead *does* lead to the feedback now, although my schematic still shows them reversed. Unfortunately, it's also very hard to see this; I attached another picture with this highlighted. I have a sneaking suspicion that having these reversed for a while may have caused some damage to the tubes. One of the tubes is melted for sure; it's internally shorted.

So, my plan at this point is:

• Tear out everything inside the amp. I have spares for almost everything except the coupling capacitors (because these are so expensive), so most components will be fresh.
• Rebuild the amp with all leads as short as possible and eliminate and crossing of leads.
• Re-check EFB voltage using a 15k resistor to ground.
• Assuming the EFB voltage is correct, install a new set of EL84s and try again.

It'll be a few days because of my work schedule, but I'm looking forward to tearing this apart and having a fresh start. I've done a lot of electronics building in the past, but for the most part, it's been almost entirely digital. One thing this project has really showed me is that analog is a lot of fun. There are a whole new set of challenges when you're concerned with more than just making sure 1s and 0s get from point A to point B. Once this amp is finally up and running, I think I might try putting together a phono preamp next.

I would drill some holes in the chassis and install some tag strips. That helps reduce the length of the leads and helps for a more organized layout.

 

[markmcmuz]

Here is my 9300 that I'm working on.

 

[Cameron06]

@markmcmuz If you're tired of waiting for your transformers, I'm not sure Edcor is the best choice. Their website says they have a 10-12 week lead time, and mine shipped at almost exactly 12 weeks.

Your 9300 looks very clean. Looks like you have a spare off in the corner, too. As of yesterday, mine is completely disassembled and ready for round 2. I did have a few tag strips before, but I think you're right, I could probably do well adding a few more. Fortunately, when I first fixed up the chassis, I drilled out every rivet and replaced them all with stainless steel machine screws, so I have about 25-30 different places I could install them without drilling extra holes. The chassis has a thin layer of body filler, primer, and paint, so drilling new holes would be a bad idea.

 

[dcgillespie]

Cameron -- In the (very nice) schematics you drew and presented in your post #41, I noticed that it did not include the EFB™ Diode Protection Pack that I now strongly recommend be included in any EFB installation using the LM337 device in the cathode circuit of the output tubes to provide EFB operation. If you use a slow warming GZ34 rectifier tube as I did in my original thread on these amplifiers, then there is less concern for including the protection package, while much greater concern if using a fast warming one like the 5U4. In either event however, I would strongly recommend you include the protection package as, while the very vast majority of LM337 examples have operated for years now (over a decade) without any concern, a few (about 8 I am aware of over a 13 year period) ultimately rebelled over the startup conditions created when power supply voltages are present well before the output tubes begin conducting current. The diode pack completely eliminates the concern to prevent any "mysterious" regulator failure. If you missed it, you can find information on it in post #1294 in my original thread here: https://audiokarma.org/forums/index.php?threads/more-fun-with-magnavox-the-9300-series.687735/

Good luck with your project!

Dave

 

[Cameron06]

@dcgillespie Thanks for looking over my schematics and for the suggestion about the diode pack. @gadget73 mentioned that to me a few days ago, so while it didn't make it into my schematic, I did install it in the final design. I've since torn down the amp to rebuild it a little more logically, but I'll definitely include the diode pack again when it put it back together (I'm using a 5U4GB).

I don't think the regulator failures I had were necessarily caused by the slow startup, though. Because of reasons still unknown to me, the current draw on the output tubes would run out of control, which eventually resulted in shorting one of them internally. The last time I blew a regulator, I observed an almost immediate 5V across the 10-ohm cathode resistor, so I'm sure I exceeded the current capacity of the LM337 (that tube is now a permanent desk ornament).

Hopefully I will have more success the second time around.

 

[Cameron06]

@gadget73

You mentioned a few days ago that I had the leads swapped on my OPT. As I've been re-assembling the amp, I noticed that the diagram from Edcor and your diagram appear to have the plate leads reversed. Yours shows the brown lead in phase with the green 16-ohm lead, while Edcor shows the blue lead in phase with the 16-ohm lead. Did you find that Edcor's schematic was wrong?

 

[gadget73]

Honestly I don't remember at this point whether I determined it was backwards or if it was marked that way. Its entirely possible they reversed the colors since I bought mine in order to match up with the Dynaco transformer its basically aimed at being a sub for. Id go with what they say unless it turns out to be wrong.

 

[Cameron06]

I hooked up my OPT to an o-scope and 5kHz signal source today to confirm: Your schematic is correct, and Edcor's schematic is backwards. I don't know if they didn't expect anyone to use the 16-ohm line for feedback or if it was just an oversight, but it's definitely frustrating. Both my of OPTs are wired that way, too.

Edit:
I sent a note to Edcor; hopefully they'll update their schematics so nobody else has this problem.

@markmcmuz - Something to consider once your Edcor transformers come in.

 

[Cameron06]

I finished rebuilding the amp. Everything inside the chassis is brand new, everything has been re-wired, all the leads are as short as possible, and I replaced all 4 output tubes.

The good news: I injected a signal, and I'm getting a clean output. My only signal source is a square wave output, so I'm getting a little bit of distortion on the leading edge, I haven't hooked it up to and audio source yet (don't have on on my workbench), but I have every reason to believe it's working. When I get some time tomorrow, I'll drag out an old set of speakers and test it.

The bad news: I *still* can't get the cathode current where it needs to be. I can adjust the bias voltage up to around 17-18V with no tubes installed. With all the tubes installed, the voltage across each 10-ohm resistor climbs to about .3V and stays there. I can adjust this upward a little bit, but I can't get it lower than .3V. My tubes are matched at 17mA, which sits right between the 22mA from Dave's schematic and the 15mA from yours, so I'm hoping to get a .17V drop across the resistor.

@gadget73 @dcgillespie I think the next step is to add a resistor to the EFB circuit in series with the 15k resistor to ground (again). Ideally, I'd like to set it so when the pot is centered, I get around 20mA at the cathodes. I really have no idea how to calculate what value I need, though. My first instinct is to add a 1k or maybe a 5k, but this could be way off. Do you think this is the right approach? If so, what should I be looking for as far as resistance?

 

[gadget73]

What sort of plate and screen voltages are you getting?


honestly just sticking a 5k in there to see what it does will probably be as effective as anything else. If these are Russian tubes, they tend to want more negative voltage than old production so trying to plug numbers into a curve calculator may not actually get us anyplace useful.

 

[Cameron06]

I couldn't find a 5k, so I used a 4.7k. That put it right where it needs to be. These are not old Russian tubes; that're Genalex Gold Lion, brand new. Regardless, these are the values I settled on for each tube:
1 - 16.5mA
2 - 16.2mA
3 - 15.2mA
4 - 24.2mA
I'm a bit annoyed that I ordered these as a matched set, and they are most definitely *not* matched. 1-3 are close enough, but I'm not too happy with #4. I thought maybe I made a mistake in the wiring, so I tried swapping them around, and sure enough, it's the tube. I'm not overly concerned, though. I'll run it this way until a tube burns up, then get a new matched set (from somewhere else this time).

Since you asked, these are the HV readings: (pin numbers included because I'm an amateur when it comes to tubes)
Screen (pin 9) - 375V
Plate (pin 7) - 375V
B+ - 378V

When I first tried the amp with an actual audio source, I noticed that the right channel was much quieter and sounded almost "anemic." No bass at all, almost like a pair of really crappy headphones. I tried swapping the power tubes, swapping the preamp tubes, switching the buffer tube, went over the wiring a dozen times... couldn't fix the problem. After more hours than I'd like to admit, I finally just plugged my phone into the input to cut out the rest of my audio system... and it sounded perfect. Turns out that there's something wrong with the right channel on my phono preamp. I don't know if my old amp had some kind of built-in channel balancing or if the preamp just happened to fail right when I tried my newly-built amplifier. So of course now I have something else to fix.

At this point, with the exception of the mismatched tubes, I'm calling this project "done" for now, unless you suggest otherwise. I really can't thank you enough for answering my questions; if it wasn't for your help, I'd still probably be trying to figure out why the EFB was so far off and most likely would have melted another set of tubes.

Pictures of the finished product are attached.

 

[gadget73]

They're still Russian tubes, and apparently share that "wants more bias voltage" thing with the Electro-Harmonix ones i have. They came out of the same factory, so not terribly surprising. What do you get from one of the bias test points to ground?

Plate and screen look right so the main supply ought to be happy.

Phono preamps have a significant EQ curve, can't use any other source with those or it sounds funny. Shouldn't be unbalanced left to right but the frequency response will be way off. Usually will have buckets of excess gain too. Phono cartridge is good for like 4mv of output, most phones will do 100mv or better. They tend to be on the weak side compared to say a CD player but still a lot more snot than a cartridge has.

Anyway glad to hear you got it sorted out. Enjoy

 

[Cameron06]

I'll re-check from the bias point to ground soon; for the time being, I've buttoned up the bottom of the amp and hooked it up to the rest of my system, so it will be a little while. I'm going to give it about 24-48 hours of burn-in time, then open it back up and re-check everything.

Regarding the phono preamp, I probably wasn't too clear about that. I'm very familiar with the RIAA curve; the preamp is connected to my turntable, which has no built-in preamp. My phono cartridge is a 5mV cartridge, and the preamp has adjustable gain. I haven't tested the output level, but I haven't had any issues with it before now. I tried moving around the tubes in the preamp (it uses 3 12AX7s), and now the right channel is even worse. If you've ever plugged a turntable directly into a power amp without a preamp, that's what it sounds like. Swapping the left/right channels on my turntable and swapping the left/right channels on my Magnavox amp prove that it's definitely a problem with the right preamp channel.

The preamp was the last piece of Amazon "Chi-Fi" equipment in my system, so I ordered a new Pro-Ject Tube Box S2 to replace it. My old Little Bear T11 does have a very nice chassis, so it will make for a great starting point to build my own preamp in the future... once I've recovered financially from buying the new one. That will leave me with: Pro-Ject DC Evo -> Pro-Ject Tube Box S2 -> The Magnavox Amp I just finished -> A pair of Polk Ti500s.

 

[gadget73]

ah, ok, I thought you were running the phone through the phono preamp. Misunderstood what you meant.

 

[Cameron06]

@gadget73 Have you had any temperature issues with your OPTs being in such close proximity to the tubes? Since the Edcor OPTs are so much larger, they sit a bit closer to the tubes than the original transformers did.

I ran the amp for several hours tonight, and it gets *very* hot. The output transformers get very warm, but not too hot to touch. The power transformer is a little warmer. The tubes are extremely hot (I burned the #$%^ out of myself on the rectifier when I accidentally laid my forearm on top of it), and I'm wondering if a heatsink or a shield between the tubes and transformers might be beneficial. I'm using slightly larger power supply capacitors, so I would expect they'd pull more current from the rectifier, but wow...

I'm planning to pick up a temperature probe just to see exactly what's going on.

 

[gadget73]

Things get warm, but thats just life around things that glow red. Haven't had any paint damage om the transformers though, so I figure its not warm enough to be a problem. Don't think I've ever measured it. Power transformer on mine doesn't get excessively hot that I recall either, though its been a little while since I ran it for long periods. Not for any special reason, just using different amps at the moment.


The tubes get hot enough to remove fingerprints. Bigger tubes are even worse. The probes on my old Simpson meter have flat spots melted in them from getting too close to the KT120 tubes in my Bogen amps.