early Fisher 400 Restoration blew a tube : (

I know this was started in an older thread but I started the restoration last night and ran into some major problems. This is an early version and was wroking fine. I am using the metalbone kit and completed the first part.
-Per my pics what is the yellow jumper wire for by the big yellow orange bias cap ? It connects to a black wire second pin in from left to right and to a side of an AC convience outlet. With the jumper removed the unit will not turn it on.
see pics here on page 2-3 the interenals are posted.
http://www.audiokarma.org/forums/showthread.php?t=133427

Now I replaced the bias circuit new caps and bridge. 123V L-N input. Prior to replacement I had -14V on V13,14,15,16 pin #2. After replacement, V13,14,15 had -16.75V and V16 -15.75V on pin #2. -27.5V out of new bridge. 423V on + side of 100uF cap C 76 and 208V on - side. Few miutes into test tube V14 blew many colors I unplugged ASAP and tried to remove but the bottom glass is broke all the way around. Needless to say this was an original Fisher tube looked on internet and original are very hard to come by and cost $$$$$$
Unit was upside down with 8ohm load attatched to outputs.
-What went wrong ? Any ideas ?
-Can I remove all tubes and test without them in and without hurting other things ?
-How do I test my output tranformers and see if they are ok ?
-My new plan is to complete the restoration with no tubes in and then put in aftermarket tubes in any ideas on what to get ? and what do I test/check prior to tube install so I do not destroy them ?

I was doing this restoration to save on parts namely tubes so far not good.:tears:
thanks in advance
Jim

-15 to -16 volts bias isn't enough, IME.

I would do the following, per advice from several people both here and elsewhere, and by PERSONAL EXPERIENCE of it WORKING GREAT:

-Replace the coupling caps, from the driver 12AX7 tubes, feeding pin 2 of the 7868s (through a 1K snubber resistor, one per tube), with .1uf 630v film caps (poly is best).

-Replace the 330K bias resistors (from the bias supply to the junction feeding the 1K resistor which is wired to pin 2 of each tube) with 220K resistors. I like using a 1 watt resistor for this, just for safety.

-Replace the 1K resistor in the bias supply (feeding the bias from the 27v supply to the bias supply lines to the output tubes) with a 820 ohm, 1 watt resistor.

This will increase the bias voltage, probably to around -19 volts. This, IME , is enough for any 7868 tubes to operate comfortably.

As for replacement tubes- I used the Electro-Harmonix 7868EH in the last one I worked on. Sounded great, completely stable, runs fine for hours without overheating...

Regards,
Gordon.

Further info:

As far as testing output transformers- check impedance from the plate leads of the output tubes (where the transformer leads connect) to the center tap of the primary (where the transformer leads connect to the B+ line). Should be about the same impedance (plus or minus 10 percent or so) on all 4 leads (2 per transformer). Should be somewhere between 50 ohms and several hundred ohms (can't remember exact value, sorry), per lead.

Also, for a more definitive test- disconnect the transformer primary center-tap from B+, and measure from that lead TO GROUND. Should be an open circuit.

If these conditions are met, the output transformers are probably OK.

You will also want to replace the bias rectifier- over near the bias cap. A good bridge rectifier (2 amp or greater) will work fine. I would recommend a good 5 or 10 amp... available almost anywhere (including Radio Shack, probably). The bias cap (the big tan cap) also needs to be replaced... just duplicate the voltage and value (in microfarads)... it doesn't matter if the new cap is a different size, physically.

There's also a 100uf aluminum-color cap next to the big tan bias supply cap... that ALSO needs to be replaced. That's the final bias supply filter cap- if that goes bad, then you have near-instantaneous melt-down of the outputs...

Regards,
Gordon.

Okay it's about time someone ends the insanity.....

These mods to the negative grid voltage are simply band-aids....kind of like operating in the dark. Blindly applying -19V or -21V or whatever your favorite scheme to get your output tubes to run safely is totally variable for each set of tubes or for that matter each tube to some degree.

The best course of action for a fixed bias amplifier like the Fishers is to first and foremost install 10 Ohm cathode resistors. So you can measure the current of each output tube and then adjust the negative control grid voltage to achieve a maximum current of 35mA for any of the 4 output tubes. Ideally they should all be within 5ma of each other with no single tube exceeding 35ma.

To achieve this the negative voltage applied to the control grid could end up being -15 or -22 depending on the set of output tubes you use. This voltage is variable and IS NOT THE BIAS. This is the control grid voltage. You measure and adjust the current of the tube. The important measurement is the current not what control grid voltage that is required to achieve the proper current.

Blindly applying -21 to a set of output tubes could end of with them conducting less then 20ma of current per tube. Yes this will make them live a very long happy life. But they will be running near class B operation. Any healthy 7868 should live absolutely happy life at 35 ma idle current. If it doesn't then it has no business being used in the amplifier in the first place since it's defective.

You know what really kills me is these amps ran wonderfully for decades without major issues and in many cases the original output tubes are still good!! What the heck makes people think they need these extreme measures now? Some mild enhancements yes but many folks seem to be going off the deep end if you ask me.


The math.... both the 7868 and 7591 have a Idle plate dissipation rating of 19 W. The equation for figuring the plate dissipation is current times plate voltage.

400V (plate) X .035 (35mA) = 14 watts...... 25% under maximum is a comfy life.

The only difference today from yester year is the tubes available back when these units were built had much tighter tolerances. Most of NOS tubes available today are most likely rejects from yester year or used pulls that have no business being used anyway. Tube testers are horrible at detecting output tube problems beyond blatantly bad tubes. The modern production versions just do not have the tight tolerance so some care needs to be taken when using them. The 10 ohm cathode resistors allow you to measure and then adjust the output section for proper operation.

Craig

" Most of NOS tubes available today are most likely rejects from yester year or used pulls that have no business being used anyway. "

I love it...and folks are paying top dollar...

I knew about the pulls part from watch TV repair men swap tubes til they found one that worked ok. I used to ask why are they putting junky tubes back in the boxes and remember being told that it might work OK in another set.

Craig thanks for the great input. I am a newbe at this tube stuff and I have some of the same thoughts you do. Did Fisher engineers not know what they were doing ??? I think they knew what they were doing and designed a great system. Currently I have redone the Bias circuit and replaced C78A, C78B, C74, with new caps put in a Si bridge ..The only mod I did was change R102 to 2.7k ohm vs 5.6k ohm. Prior to power up. Wish I had a tube tester. I just do not want to hurt anything in the system while doing checks. I blew one tube already do not even know why it died. Perhaps it was its time.
Is it a problem running the system upside down ???????

PER your post R102 may have to change to get @ 35mA of Bias current wich will be measured off the new 10 ohm resistors. The Bias voltage can be between -15 and -20V measured off of pin #2 or 6 on output tubes.



I want a way to test the output transformers without tubes to make sure they are good to go before I put new tubes in. I will try what Gordon sugests

I Completed B+ Cap change and diode change last night.
Now I am putting in the the 10 ohm resistors on Pin #3 and will be replacing the caps in the Amplifier Signal Path with new caps keeping the same values.

What peoples thoughts on EH output tubes. I am thinking of getting a matched quad and taking the rest of the original out until the refurb is done and make sure it is running stable.
These pics were before I started retoration. I will try to upload what I have done later.

Also what is the yellow jumper wire for it will not power up without it ??

It's a simple FACT that the greatest allowable grid bias resistor, for fixed bias on either 7868 tubes OR 7591 tubes, is 300K.

Fisher uses 330K... which was PUSHING IT, even with GOOD NOS tubes.

Unless you're going to discount the original manuals, AND the experience of very knowledgable people (not me- Jim McShane is the principal one who originally suggested this- I just found out that it WORKS- and makes the amp work PROPERLY), then changing the coupling caps and the bias supply resistors to something like 220K ohms is PROPER SOUND ENGINEERING PRACTICE.

The change to 19 volts bias does NOT put the tube into Class B. Not really even close to it, with most tubes tested here. Only AFTER changing the bias to around 19 volts, could I get the vast majority of tubes to test between 30 and 40ma of bias. REMEMBER- we're dealing with 120 TO 125 VOLTS LINE VOLTAGE in most cases in most metropolitan areas, not 110V! A bias voltage which would result in 35ma at 117V, will usually wind up idling at 45ma or so (or more), at 125V! It's not a linear relationship- the bias current goes up MUCH FASTER than the increase in B+ voltage for the same bias voltage!

And as many blown Fisher 400 output transformers as I've been hearing about, from VARIED sources- a bit of extra care seems WARRANTED in this day and age, where you can't just go out and buy another one off-the-shelf...

So, if you wanna push it to the limit, feel free... it's your amp. If you want it to actually LIVE under modern line voltage, you MIGHT think about doing what I suggested...

Regards,
Gordon.

Oh, BTW: On 110V, the Fisher 400 is SUPPOSED to have -17v negative grid voltage on the output tubes, with about 410 volts on the B+... STOCK. This is from the FACTORY SERVICE MANUAL. I have one of those... I can actually look it up. Funny that.

Under 125V, that scales to about -19v, (and 435V or so B+) WITHOUT "MODS" to the operating point of the circuit. All we're doing here, is PRESERVING the same relationships, at a minimum!!

NO wonder the grid voltage has to be increased to -19v, then, huh? Even increasing it to 20V, many tubes STILL are passing more than 30ma at idle, from my tests here...

Regards,
Gordon.

Gordan.......

Fallacy #1 no tube gear built in the 60's was designed to run at 110V look at the schematics almost all will state 117V.

Fallacy #2 look at the specs on a 7868 or 7591 and find this grid bias resistor information?? If indeed Fisher, Scott and countless other engineers were pushing this specification then why the heck did the tubes last so long???? I guess these engineers were just a bunch of idiots and you no better Huh!!!! In reality lowering the grid resistor value is just a means to bring the negative grid voltage more negative (with known good tubes changing from 330k to 220K isn't really going to do all that much). The value of the resistor matters very little as long as the grid voltage is sufficient to set the proper current for the tube in question. The only other benefit of the lower value grid resistor would be if a person was using tubes with grid leakage the lower value resistor would allow less voltage drop as the negative grid voltage was being sucked from the supply from a defective tube. I say if your using tubes with grid leakage your asking for trouble the tube is defective throw it away.


Fallacy #3 where did I say that you should not or could not end up at -19V in some situations, with some tubes and some wall voltages??? I said that no absolute grid voltage is right for every situation. I stated it may end up -22V and I've actually seen sets of 7591 tubes require that much in some vintage amplifiers. I never said that the tubes WOULD be at 20mA I said they COULD be........ big difference. Again checking the proper pertinent Cathode voltage will insure the receiver is biased safely. Guessing with some magical grid voltage number is much more careless if you ask me.

I said to check bias on the 10 ohm resistor and adjust the negative grid voltage as required to get the tube running at the proper current. The NEGATIVE VOLTAGE IS VARIABLE and IS ABSOLUTELY NOT THE IMPORTANT VOLTAGE TO MEASURE. Your comments about wall voltage, B+ voltage and Negative grid voltage are exactly what I'm talking about. THESE ARE ALL VARIABLES. The 10 Ohm cathode resistors take the unknown variables out of the picture for each person working on one of these projects. I myself install a bias pot under the chassis to adjust this negative grid voltage up and down. I have never out of 100's of these amps and receivers ending up setting the negative grid voltage exactly the same to achieve proper current. It's always slightly different and some times drastically different one receiver/set of tubes to the next... Especially if switching to the modern production replacements.

Why do you suppose output tubes are sold in matched pairs, quads and so on?? I can tell you first hand..... because they are not built to tight tolerances even back in the day they were not perfect but much better then what available NOS or modern production today.

Have you ever purchased and matched a bulk of output tubes? The current from the lowest to the highest can be pretty extreme out of a lot of 100. If it wasn't the case they wouldn't need to match them... it would be a huge waste of time and resources.

My beef is with folks suggesting that some magical grid voltage number exists. It does not. The 10 Ohm cathode resistor measurement is the truly important thing here. How you go about adjusting the applied negative grid voltage is up to you. Many way exist to accomplish it. I have never change a grid resistor value myself because it simply just does not need to be done to accomplish the job at hand.


Craig

NOSValves said:

I myself install a bias pot under the chassis to adjust this negative grid voltage up and down.

Craig

Click to expand...

Craig, if you don't mind me asking, what resistance pot do you use? The fixed bias design bothers me.

Speaking for myself, you'lls debate is very insightful for us grass hoppers around here.

Gentlemen, I appreciate the discussion and don't want to shut anyone down because it's enlightening but hopefully we all keep in mind that we're FRIENDS discussing our hobby.

Whaleman said:

Craig, if you don't mind me asking, what resistance pot do you use? The fixed bias design bothers me.

Speaking for myself, you'lls debate is very insightful for us grass hoppers around here.

Click to expand...

In this case where the existing R102 was 5.6k 1/2 watt I would use a 10K 1/2 watt pot in it's place. Just pre-adjust it for about 5K then do the final adjustment when you fire up the receiver and then again after it has warmed up for a couple hours.

Some of these receivers (500B) actually have a 10K resistor at this point and in that case I would use a 20K pot. Again make a pre install adjustment so the value is around the original resistor value.

These pots have 3 solder lugs. I solder one of the outer lugs to the wiper (center lug) making this effectively one lug so now you have 2 connections just like the existing resistor. You solder these lug to where the leads of the resistor were sodered (remove the orignal resistor of course).

This resistor to ground is the easiest method to adjust the grid voltage. Replace the resistor with a pot and you can adjust the grid voltage up and down. I'd never recommend installing this type of control on top of the chassis though and also caution you to adjust it with care. This control can take the grid voltage down to zero if adjusted improperly which would not be a good thing.

Craig

NOSValves said:

The best course of action for a fixed bias amplifier like the Fishers is to first and foremost install 10 Ohm cathode resistors. So you can measure the current of each output tube and then adjust the negative control grid voltage to achieve a maximum current of 35mA for any of the 4 output tubes. Ideally they should all be within 5ma of each other with no single tube exceeding 35ma.

Click to expand...

OK - Great discussion here. Since I have read (sorry Craig - still reading) - many differing opinions on this topic, maybe we can get some agreement on what the current should be of the output tube. As Craig says everyone has a differing opinions of what the voltage should be. If it is 35 ma, then for my Fisher 500B at my home the bias voltage is -15.5 volts.

Cut-Throat said:

OK - Great discussion here. Since I have read (sorry Craig - still reading) - many differing opinions on this topic, maybe we can get some agreement on what the current should be of the output tube. As Craig says everyone has a differing opinions of what the voltage should be. If it is 35 ma, then for my Fisher 500B at my home the bias voltage is -15.5 volts.

Click to expand...

I think this is the best post I have seen for awhile on this topic and I have a lot of usful information to ponder through and some questions. trying to use logic:
-I think it is best to vary the Bais circuit vs making changes to the Amp stage. It makes more sence to me to alter what is suppose to give the output tubes there bias and not to alter someting else to compensate it.
-35mA max is a constant and putting in the 10 ohm resistors make more sence to me then anything .....They offer protection.....One can easily find out if the tube is bais properly. What a great idea
QUESTION:
If one alters the Bias voltage -15--20 or whatever is needed to bias to the the output tubes is there any risk of this affecting anything else.....ie the Preamp tubes or other parts of the 400 where the bias is applied ?
(I need to look at the schematic in more detail to see this feeds anyhting else)

Thank YOU ALL for the GREAT INPUT on this topic. This is truly a great forum where minds though may be different in approach come together.
There is always more then one way to skin a cat.
Thanks again.
Jim

bowtieman427 said:

I
QUESTION:
If one alters the Bias voltage -15--20 or whatever is needed to bias to the the output tubes is there any risk of this affecting anything else.....ie the Preamp tubes or other parts of the 400 where the bias is applied ?
(I need to look at the schematic in more detail to see this feeds anyhting else)

Jim

Click to expand...

Changing the grid voltage, itself, won't affect the 12AX7. It's protected by the coupling cap. It won't know the difference, in short.

Changing the coupling cap values and the bias-supply resistors (to .1uf and 220K respectively, as recommended by Jim McShane and others) WILL change the load and drive current of the 12AX7 a bit... but it still WELL WITHIN the proper operating limit range for anode load on a 12AX7 in a circuit like this. You will still have PLENTY of preamp drive... the preamp will still have "headroom" when the outputs get to clipping. I've verified this in-circuit, as have others.

In short- if you WANT to install the separate bias pots and such as Craig recommends, that's absolutely fine. If the amp is a keeper, and you don't mind having parts not in the original locations hanging around in the amp, then yes, separate bias pots FOR EACH TUBE is the way to do it most accurately. For those who want to keep their amps "stock appearance", what I suggested IS SAFE and WILL SOUND FINE.

The only HAZARDOUS route is to leave it STOCK, and expect it to reliably work on modern line voltages. Even the increase from 117v to the (seen in MOST metropolitan areas) 125 or so volts on the AC wall outlet, will cause the working point of the tube to change, to a point where it's perilously close (if not over) the safe operating limit for SOME tubes. What I suggested WILL WORK for the VAST MAJORITY of tubes out there... it's been tested extensively, not just by me, but by others...

Oh- and as for the MANUFACTURER SPECIFICATION of not wanting more than 300K on the grid bias-supply resistor:

http://www.pmillett.com/tubedata/HB-3/Receiving_Tubes_Part_2/7868.PDF

I quote:

Maximum-Circuit Values:
Grid-No.-1 Circuit Resistance:
For fixed-bias operation. . . . . . . . . . . 0.3 megaohm

For a reasonable margin of safety (25%), 220K ohm is a QUITE REASONABLE value. Yes, you could go to 250K or 277K... it SHOULD still work fine. But, 220K is quite fine for the 12AX7s, and is a more commonly available value...

Regards,
Gordon.

You found 300K wow I looked at a few tube specs sheets trying to find it.... I still say these receivers worked fine as is for 5 decades with the 330K no big deal. Some mild changes are in order no matter what way you look at it. The pot is my nifty simple way to solve the issue and make it easily servicable for future tube changes. One does not have to install the pot they can simple vary this resistor value if they don't want a pot under the chassis.

To me changing 9 part values versus added 4 resistors and changing the value of one part both methods change the circuit neither method is stock. If you ask me the 9 parts value changes is further from stock. I ask which makes the receiver more versatile and user freindly? Which method raises the safety margin in the event of tube failure and possible transformer failure? Which method keeps the original sonic character as Fisher designed it?


In essense your method is just masking a tube problem. I bet in every single case just raising the grid voltage will fix the problem associated to hot biasing tubes. If it doesn't then the tube have grid leakage which is causing voltage drop across the 330K resistor. The 220K resistor may afford less voltage drop but the tubes are still defective and it will get worse with age.

Grid leakage = defective tube.

Craig

While were at this we might just as well open up another can of worms.

When changing out the selenium rectifier the 12AX7's preamp and phono stage negative heater voltage is also raised above specification. Which will indeed change the sonic character of the 12AX7 and stress the heaters on these tubes. The spec on the schematic is -24V (makes sense since these tubes are 12V heaters and two pairs are run in series) this voltage usually ends up to about -27V. I've never read anything about fixing this out of spec condition on the internet. Fairly simple way to fix it is to add a resistor and another filter cap 10 to 15 ohm resistor and 100uf 50V cap. This resistor needs to be 3 watt mimimum. This extra step before the heaters will also lower the residue ripple in the heater supply which is a minor improvement.

Craig

IMO, it sounds like that 7868 was already cracked, and the fireworks show happened when you applied voltage to a tube full of air.

fsjonsey said:

IMO, it sounds like that 7868 was already cracked, and the fireworks show happened when you applied voltage to a tube full of air.

Click to expand...

I hope you are right. It may have been the case since it has been moved around a little. I checked the output transformer per what Gordon posted and they were 114, 106, 114, & 106 ohms from tranformer outs to tubes to center tap. Hopfully this is acceptable.

Craig, any points per early 400 schematic I should put in the RC network of 10-15 ohms 3 W and 100uf 50V cap I am assuming the 10-15 is in series and the cap would be (most likey be polar) the + would go to ground because of negative Voltage.
Thanks for the input
Jim

The resistor and cap can be installed any place you can make them fit neatly. I do it right near the voltage source at the existing RC network. If you look at the schematic you will see the connection that heads off to the 12AX7
heaters I believe in the 400 this wire is yellow but my memery isn't all that good model to model. In the 500B it's purple. Yes the 100uF cap is an polar electrolytic and the positive lead goes to ground since your dealing with negative voltage.

Your measurement on the transformers are fine

Craig