Magnavox Imperial model 1ST233

[electra225]

Perhaps. Bill's Imperial is the standard we are judging this one by. His works like it should. I am comparing considering differences between his and mine. If I plug up the holes and make it worse (I can't imagine it being WORSE) I can return it to original. If Magnavox plugged up the holes in Bill's stereo, they did that for a reason.

[William]

Just so you know, Chris, my Imperial has the holes covered right from the factory and mine will pound nails for all the bass it has. From research, mine is one of the last Big Imperials built by Magnavox before switching totally to Astro-Sonic. I'm guessing Magnavox had complaints from customers to why their very expensive console stereos were weak in bass. Or maybe they were trying to compete with components, or we may never know the real reason why the older ones were weak in the bass department and mine is not. Once Astro-Sonic came out, Magnavox stereos had plenty of bass.

Bill

[electra225]

I have found several of the resistors I didn't replace last time have drifted high. The ones of particular interest are the cathode bias resistors in the 12AT7 audio amplifiers. Using Uncle Doug's formula, the crossover on the bass amp, the frequency where 100% of frequencies are passed is around 6600 hz. This almost totally attenuates the bass frequencies, making the bass amps midrange amps, there are no bass frequencies being passed 100% to the bass amps. I changed the 220K grid resistors last time, those are still within the 15% tolerance specified on the schematic. If I get the resistors values right, then change the cathode bypass caps on the 12AT7, the crossover should then be down about 200hz, with 45 decibel gain. The current value listed now is .1uf. The plan would be to change the value to 1.0 uf. The 2200 ohm bias resistors have drifted, one to 3830 ohms, the other to 4200 ohms, both above the 15% tolerance.

[electra225]

I read on the AES site where guitar guys use a 12AT7 in place of a 12AX7 to reduce gain. I wonder what would happen if I replaced the first audio tube in the amp, a 12AT7 with a 12AX7 to increase gain. The treble amp uses a 12AX7 as first audio. I have that on my list of things to try when the time comes. I know that if I use a 12AX7 in place of a 12AT7, the frequencies passed to the bass amp will be different.

The main difference between Bill's Imperial amp and mine is the power transformer. His runs 350 volts on the plates of 6V6's. Mine runs 320 volts on the plates. The maximum plate voltage listed in the RCA tube manual is 315 for single-ended, 295 for push-pull. Why is Magnavox running the output tube so hot? Bills has essentially the same cathode components and grid resistors as mine, so his is biased hotter yet? I watched one Uncle Doug video that said some guitar amps run 6V6's to 400 volts on the plates. Somebody smarter than me may understand how this works.

[electra225]

I have the resistance measurements done. If I don't measure the wrong tube pin (?!) I haven't found any huge areas of concern. I found R54, a 680K grid bias resistor in V12 a bass channel audio amp that had drifted. The 2.2K bias resistor was okay. In the other 12AT7 audio tube, just the opposite was true. The grid bias resistor was okay, but the 5600 ohm inverter resistor and the cathode bias resistor had both drifted high. I am going back over my work and my calculations. I'm using a 10% tolerance rather than 15% because it's easier to do the math in my head. There are several 220K ohm and 22K ohm resistors that have drifted high. I'm on the fence about totally recapping the chassis with the exception of the power supply electrolytics. I see I didn't put a fuse in the primary of the PT, so I'll have to do that. I used the "little cheap yellow" coupling caps. I wonder if there would be a benefit to using "better quality" caps to replace these. This amplifier uses the same cathode biasing components for all four bass amp output tubes. I wonder if it would be better to split this up to just two tubes and run another set of biasing components for the other set of output tubes.

[William]

I find it interesting, when I order new resistor, getting them from Mike, they usually arrive either 2% or 5% tolerance. That really makes things tight, which I guess is good?? I asked him why the tight tolerance. His reply was if I remember to order them old school, they will come like that. Otherwise, they automatically come the tight tolerance.

Just for information, when I overhauled my Imperial, I used the yellow caps because that is what Sal was selling. Having said that, I'm not sure if more expensive caps could possibly make mine sound any better.

Bill

[electra225]

The 81-series amp output tubes used on set of bias components for all four tubes. If I know the plate dissipation and if I know the plate voltage, and if I know the current plate current, should I not be able to reverse Ohm's Law and figure out what values I would need to split the two pairs of output tubes and put each pair on their own bias components? I wonder if the Concert Grand has all six output tubes in each amp all connected together, sharing bias components. Seems to me that this is a bean-counter move.

[electra225]

I'm not an expert on the nitty-gritty of amplifiers, the finer points of operation. Biasing and plate current is not something I have worried about before. One has to be careful of what one reads on the Internet. AudioKarma is little help. They are concerned about "hearing" rectifiers, having black plates and other nonsense. What I have been trying to watch are videos of known competent practitioners, such as Uncle Doug, who genuinely understand how this works and are articulate enough to teach others.

I've learned that having correct-value components, particularly resistors, and to have those components the same value between channels is key. The resistor values in this amp are all over the place. There is considerable difference between channels. I'll probably end up replacing all the resistors I didn't replace last time. I've read where using close to exactly-specified resistors can cause something to not work properly. I'm not sure I hold to that theory. If an exact value was unimportant, why do they specify an exact value? Why not say that a value from "here" to "here" will work? I can get 1% tolerance resistors. Then make sure they match as close as possible between channels. The three most important resistors are the grid bias, cathode bias and plate load resistors. The values of those vary widely between channels in this amp. I don't know if this will make a difference in performance or not. I want to have the right value components, matched as much as practical, and have a solid baseline to start.

The step after that will be to make the biasing measurements. After I get the biasing like it needs to be, we'll see what we have. I don't know how robust the output transformers are as far as current passing thru them is concerned. Magnavox has been criticized for having wimpy output transformers as far as performance is concerned, but I don't know how stout they are for current. My guess, based on what I have seen so far, is that these bass amps are biased fairly "cold" with low plate dissipation. I'll answer this question for sure in the biasing procedure. I'd like to run as close to the 12 watts of plate dissipation that 6V6's are rated for as I can. This will give me better fidelity and more headroom. I hope the transformers will cooperate. If not, we have answered the "wimpy output transformer" question for sure.

[Conelrad]

The discussion about 'sealing' up the cabinet hasn't mentioned you (Greg) would be changing the enclosure from a ported reflex to a acoustic suspension system. The big difference in them is the bass response with the proper woofer.

DG

[electra225]

I have a hunk of lauan I'm planning on stapling over the hole in the bottom of the speaker enclosure. If it helps, fine. If it doesn't help, but doesn't hurt, I can leave it or remove it. If it hurts, I'll remove it. There won't be any permanent damage to the cabinet and I won't know until I try. Magnavox closed that hole between the 1960 models and the 1962 models. My methods may not be very scientific, but I'm having fun and learning a lot that I didn't know. I am convinced that Magnavox didn't build this stereo to sound as bad as it does. I'm determined to try and find out what is wrong with it.

[walyfd]

Is there a way to bypass the internal speakers and hook up a known, good external set?

[electra225]

Well, sort of....

I have a set of speakers that came out of a solid state Stereo Theater, both the 15"s (made by Carbonneau) and the horns (unknown code). I can sub these for the original speakers. These may not be optimal, but maybe I could hear a difference between them and the originals. I am also planning to connect the Marantz 2245 in my "laboratory" (den) to the speakers in the Imperial cabinet. This to see if the original speakers are capable of reproducing bass.

[walyfd]

Last year, I got the 91 started after winter. The radio was dead. Had to send it out for tuner and amp rebuild. When I got it back and tested each speaker, I found out the right rear was dead...

Went to the junkyard and pulled 3 correct speakers. Only ONE was good.

So, it's possible your speakers are suspect. You try reversing the leads?

[electra225]

The only testing I have done at this point is to make sure the speakers were wired right, that the polarity on the speakers was correct. They appear to be wired correctly.

[William]

With the previous photos Greg provided in this post, his 15" woofer look just like the ones I have in my Imperial. Trust me when I say those 15" woofers will take and provide all the bass you want.

Bill

[electra225]

Maybe some progress? Maybe some light shed on the cause of anemic bass? There are 29 capacitors in the amp that I'm dealing with. 13 are good, four of those were replaced in 2015. The remaining 16 are from bad to really bad. Many of the most drifted are in cathode biasing duty, plate load and grid bias, the "big three" resistors in audio amps with P-P output. The cathode bypass cap for the bass amps is spec'd at 130 ohms, the one there is rated 150 ohms, measures 146.1 ohms. It will be interesting to see what effect this has on plate dissipation when time for that comes. I don't know what element voltages are at this time, so there is that as well. We know we have weak tubes in every position in the amp. I am running my measurements giving a 5% tolerance, since I want everything on spec and balanced, also due to the fact I can get 5% resistors. There are two resistors that check good, but are more than 5% in resistance difference between channels. I am running the treble channels to the same tolerances. I want to run the resistance readings thru Uncle Doug's formula to see what frequencies the bass amp is dealing with currently.

[electra225]

Oddly enough, the drifted resistors actually IMPROVE the bass frequencies input to the bass amps. The difference will be putting all the resistors to their specified value, than changing the cathode bypass cap value in the 12AT7's from .1 uf to 2.0 uf. That changes the crossover frequency to the bass amps from around 6600 hz down to 30hz. The jury is still out on whether Magnavox designed wimpy bass response into this stereo or not.

[electra225]

At the risk of getting off into the weeds here, I ran across another little tidbit of information I'm not sure I understand. I have been watching Uncle Doug videos on the different nuances of amplifiers. Uncle Doug is primarily a guitar amp guy, but the basics between guitar amps and audio amps are the same. I'm not the best "searcher" in the world, but it seems to me like this information pertaining to tube, stereo audio amps is woefully scarce. There are other videos on amplifier servicing besides Uncle Doug's, but many use slang and techno-babble that more confuses than enlightens. Uncle Doug speaks clearly, simply, and uses little hand-drawn diagrams for illustration.

Uncle Doug recommends biasing output tubes to no more than 80%-90% of their rated maximum plate dissipation value when using cathode biasing, like that found in this Magnavox 81-series amplifier. 90% of the 12 watts maximum plate dissipation rating on a 6V6 would be approximately 10.8 watts. This recommendation is made in order to extend tube life. Apparently guitar amplifiers are operated in a different world than a stereo audio amplifier would typically be. Guitar amps need gain, but distortion is not a consideration, and, in fact, is actually desirable. Some guitar amps are run with plate dissipation as high as 110% of ratings, right on the ragged edge of red-plating. Magnavox output transformers are frequently criticized for being a bit anemic. I reckon if you fry one and replace it with something more robust that would not be the worst thing in the world. Plus, Uncle Doug's videos deal with mono guitar amps, not Magnavox stereo bi-amps.

I'm trying to learn biasing and believe I may be making progress. I'm trying to steer clear of the "loony fringe" which certainly has a robust following. I stay clear of most guitar amp discussions as well as some of the more famous audio forums. I believe I have had guidance that puts me on the right path. There are a lot of gaps yet. Since one can find an old Magnavox stereo with an inch of dust on the chassis, but it will still be running its original tubes in most cases. This fact would indicate that Magnavox biases their amps in total relatively conservatively. If that is the case, perhaps we can find a happy medium where we can get maximum performance as well as acceptably long tube life. I'd sure welcome comments and suggestions or hearing about experience you guys have had in this regard.

[electra225]

I've got all the resistors changed that I had the value on hand. I have the ones on the individual channels balanced as much as I can, some within a few ohms of each other. Less than a 1% tolerance. I found that I do not have a quad of Sylvania rectifiers like I thought. I put one in the Symphony. I'll need a new pair of rectifiers, a matched pair of 6V6's and a matched quad of 6V6's. I have plenty of good 12AT7's and enough good 12AX7's if I need one. I tossed about two dozen tubes I had on hand that tested weak or flunked the life test. I'll order the resistors I need from AES, since that is where I'll buy the tubes I need. I read on the AES site where JJ 6V6's can run up to 475 volts on the plate and can go up to 14 watts plate dissipation. That sounds like running them pretty hot, hotter than the output transformers can handle. I think I'll get Electro-Harmonix 5U4's and then JJ 6V6's. I need new output tubes before I can do the biasing. I'll run another resistance test like I did initially to make sure everything is 100% before I power up with new tubes.

[electra225]

I found some mismatched "not junk, better than what was there" tubes to replace the ones in the amp chassis that were beyond use. I ordered the residtors and a couple other small parts I needed from AES, and will actually spring for shipping, since the items and their packaging is small and shipping is cheaper than going up there after them. I can't do much more until I get the parts I need to finish. After I get the resistors changed, next step will be to run the resistance checks again. We'll take it from there once it passes the resistance tests ....