The nice thing about the 8950 and the M-2057 are that these tubes give you a pair of 1 amp fusible cathode links on pins 2 and 6. You get two chances to destroy each of those tubes.We call this the "blown fuse" syndrome.
The 8950 tube seems to be susceptible to this sort of failure.
This one was.
Learned to eyeball that thin link from pin 3 to the cathode at the center of the tube before even bothering to put it into the tube tester.
73
-
You can now help support WorldwideDX when you shop on Amazon at no additional cost to you! Simply follow this Shop on Amazon link first and a portion of any purchase is sent to WorldwideDX to help with site costs.
-
A Winner has been chosen for the 2026 July 4th Retevis RA89R Giveaway! Click Here to see who won!
Super scout info needed pls
- Thread starter Cable Guy
- Start date
The glitch resistor makes a huge difference. I back that up with a cathode fuse when a single stage has a bunch of sleep tubes. One extra step when you're dealing with six tubes in a final like this, is to balance them out with one ohm non-inductive resistors feeding each cathode. The fuse handles a combined overload affecting all tubes. The 1 ohm and glitch resistors handle internal tube arcs affecting individual tubes.
Your idea to link the two cathode pins together is exactly what we do when one fails. I don't like to do this in advance, because it will allow twice as much current to flow through each tube and provides no second chance in the event of a severe overload. That's something you do after you've lost one of the two fusible links. Unfortunately, I am only aware of the 8950 and M2057, having double cathode pins and fusible links. The exception to that was National brand 20LF6's. They were the only brand of that model tube to double link the cathode and unfortunately, some amplifiers used the extra pin on the socket as a binding point and you would get fireworks plugging this brand in. The eagle 525 comes to mind as one amplifier that destroyed a batch of new National 20LF6s.
The thing to look for is the plates on the Plate Tune control. The "twin peaks" rule is that in one 360-degree rotation of the knob you should see TWO positions with a peak on the meter. They might be close together, or 180 degrees apart. What counts is seeing two distinct peak positions of the Plate Tune knob.
And if you see only one peak, it's not really a peak. A glance behind the front panel will reveal that the control's plates are either full apart of fully meshed together. The control reached its end of travel before reaching a proper resonant peak.
I'll skip the detailed discussion for now, but making sure the Plate Tune control reaches a proper resonant peak is the best way to prevent a repeat performance from the next set of final tubes. If it won't the tank coil connected to the Tune capacitor will need to be adjusted so the Tune control reaches a peak before running out of travel.
73
And if you see only one peak, it's not really a peak. A glance behind the front panel will reveal that the control's plates are either full apart of fully meshed together. The control reached its end of travel before reaching a proper resonant peak.
I'll skip the detailed discussion for now, but making sure the Plate Tune control reaches a proper resonant peak is the best way to prevent a repeat performance from the next set of final tubes. If it won't the tank coil connected to the Tune capacitor will need to be adjusted so the Tune control reaches a peak before running out of travel.
73
My friend brought 8 tubes to replace the ones he blew up. I set out to check this amp, and it does seem to work now, however, the final tune cap is completely un-meshed for a (false) peak. Could be why it blew up in the first place.
I'm going to adjust the final output coil and see if I can bring it back. Do I need to remove turns?
Edited to add, load is about half way meshed. Same for driver tune.
Last edited:
Simple rule to decide whether to decrease or increase the coil's inductance. If the plates are all the way apart, this is the control's minimum setting. A circuit with "too much". Can't make the capacitor smaller, you need to make the coil's inductance smaller. Stretching turns farther apart, or a short between two adjacent turns.
If this overdoes it, the Plate Tune will now appear to "peak" at the opposite extreme with the plates fully meshed together. That would indicate the need to push the turns closer together, or unsolder a jumper that's shorting a turn to the one next to it. Plates fully together, the circuit doesn't have enough. Could make the cap bigger, but the coil is easier to adjust.
If the plates are fully separated, reduce the coil's inductance.
If the plates are fully meshed together, increase the coil's inductance.
I aim to get the coil tweaked so the control will peak at the halfway point. Not that there is anything magic about it, but that peak position will be influenced by SWR. If the control peaks in the middle on my dummy load, there will be leeway in both directions when it's connected to an antenna.
Tweaking a pi-network coil is a bit like an artillery officer getting his barrel elevation right for range to target. First shot is short, lower the barrel. Next shot is long, raise the barrel half as far as you lowered it. If it's short now, lower it half as far as you raised it. They call it bracketing. Eventually you zero in on the target.
73
If this overdoes it, the Plate Tune will now appear to "peak" at the opposite extreme with the plates fully meshed together. That would indicate the need to push the turns closer together, or unsolder a jumper that's shorting a turn to the one next to it. Plates fully together, the circuit doesn't have enough. Could make the cap bigger, but the coil is easier to adjust.
If the plates are fully separated, reduce the coil's inductance.
If the plates are fully meshed together, increase the coil's inductance.
I aim to get the coil tweaked so the control will peak at the halfway point. Not that there is anything magic about it, but that peak position will be influenced by SWR. If the control peaks in the middle on my dummy load, there will be leeway in both directions when it's connected to an antenna.
Tweaking a pi-network coil is a bit like an artillery officer getting his barrel elevation right for range to target. First shot is short, lower the barrel. Next shot is long, raise the barrel half as far as you lowered it. If it's short now, lower it half as far as you raised it. They call it bracketing. Eventually you zero in on the target.
73
Thank you sir. I shorted a turn and it brought the tune control to approx. half mesh position peak, and there are now 2 peaks in a 360 degree rotation. It did move the load peak around some, but still between max and min capacitance.
Oh, what size fuse should I use for this 4x8? I have a 30A in it now but I don't want to oversize it. My 2x4 calls for a 15A, I assumed double for a 4x8.
Assuming it has a standard 120 plug on it, you can't have anything larger than a 20 amp breaker at the box. I would try the lowest fuse that holds reliably under normal operating conditions. That will offer the most protection if accidentally operated outside of safe conditions. I don't think I would be comfortable going past 20 amps.
I can tell one possible cause for burning out the pin link someone was over driving the amplifier. I have seen it many times. I have even did it myself
That's what I'm thinking on this also. The owners rarely admit they were overdriving it when it went poof and it was always in the back of my mind.
Update, after a few years of service the owner needed some cash and sold it, the new owner did the same thing as the last and popped all the final cathode pins, same description of a flash under the hood, suspect either improper tune or overdriven.
Fortunately, I remembered this amp and knew what to look for. Check grid bias, caps, plate voltage, new toobs, twin peaks, back in business. I'm thinking of adding cathode resistors or fuses to prevent it from happening again.
I told the guy not to use anything but a clean output, single final radio, 2-4 dead key max, no more than 12-16 pep for longevity. I warned against a very low dead key or ssb use can overheat the drivers on HI. Maybe they will listen, those toobs are getting rare and expensive and you would think the expense alone would be a good lesson, but some folks you gotta kick 'em in the head to learn 'em something.
Edit to add, thanks again for sharing your thoughts Shockwave and Nomad. I have learned much from you guys and owe much of my success to you and the other forum greats.
Fortunately, I remembered this amp and knew what to look for. Check grid bias, caps, plate voltage, new toobs, twin peaks, back in business. I'm thinking of adding cathode resistors or fuses to prevent it from happening again.
I told the guy not to use anything but a clean output, single final radio, 2-4 dead key max, no more than 12-16 pep for longevity. I warned against a very low dead key or ssb use can overheat the drivers on HI. Maybe they will listen, those toobs are getting rare and expensive and you would think the expense alone would be a good lesson, but some folks you gotta kick 'em in the head to learn 'em something.
Edit to add, thanks again for sharing your thoughts Shockwave and Nomad. I have learned much from you guys and owe much of my success to you and the other forum greats.
Last edited:
Last edited: