How do you increase headroom in a 50w JMP?

[william vogel]

Do KT88s last longer than EL34s? I ask since they cost about $130 for a pair without shipping

In theory they should last about twice as long because they have the ability to make twice the power if run at their ability. In reality, I can’t tell you. I build amps and I sell them. I’ve had power tube failures early in their life (<1 hour) but I’ve never worn out a single power tube. I think you’ll be able to get many hours of playing and much longer life than EL34’s. I have the Sino brand Chinese KT88’s in my amp right now. I’ve used the Gold Lions in a DR103 build and the Sino again in a Super Bass build. Both those amps belong to the same customer in Texas and they are 2 years old and still running strong.

 

[ELS]

In theory they should last about twice as long because they have the ability to make twice the power if run at their ability. In reality, I can’t tell you. I build amps and I sell them. I’ve had power tube failures early in their life (<1 hour) but I’ve never worn out a single power tube. I think you’ll be able to get many hours of playing and much longer life than EL34’s. I have the Sino brand Chinese KT88’s in my amp right now. I’ve used the Gold Lions in a DR103 build and the Sino again in a Super Bass build. Both those amps belong to the same customer in Texas and they are 2 years old and still running strong.

that's not really how tube life works, if you run the tube hard the cathode will start stripping it's coating off and destroy the tube quicker, if not it will still slowly evaporate.
so a tube that has twice the max plate dissipation means nothing in tube life, since the plate isn't what ages in the tube, it's the cathode.
for example an EL500 tube has a huge cathode to allow for a lot of current flow across the tube, but the plate dissipation is only like 18.5w iirc, it was designed for use in horizontal output circuits in TV's where the tube only needs to supply a short pulse of current instead of continuously conducting like in an audio amp.
if you used such a tube in an audio amp, you could even red plate the tube and it will last a long time (unless the plate melts) since the cathode was designed for a ton of emission.
however if you used like an EL34 which has a considerably smaller cathode in a horizontal output circuit, the cathode would just be nuked.

what I mean is that a larger cathode will last longer, a bigger plate tho wont.
higher plate dissipation demands a bigger cathode so it lasts at all, but the cathode wont always be matched to the max plate dissipation.

that's why for example the EL500 will red plate so hard that it melts the glass envelope; the cathode isn't limiting the current flow, nor is it stripping super fast.
while it's much less common on tubes like the EL34; cathode wont be able to emit any more electrons and will limit the current, and then before the glass envelope melts the cathode will already have stripped the coating off so much that emission will drop so much that the plate starts to cool down.

long lasting tubes would be for example:
6550; EL500; 6CB5A; 6LF6; 6LQ6; or many other horizontal output tubes.
if you look at Output-Transformer-Less amp designs, they will use tubes with huge cathodes so they can get a lower output impeadance, tubes like that will usually last way longer than normal audio amp tubes.

 

[Pete Farrington]

using an ultra-linear output transformer, would increase outupt power from ~50watts up to around 70 if you get it spot on

That seems an unusual viewpoint, runs counter to my understanding and experience; could you clarify?

 

[ELS]

That seems an unusual viewpoint, runs counter to my understanding and experience; could you clarify?

ultra linear output configuration adds negative feedback to the screen, it will increase headroom in the output. that's all I can clarify about that since I've rarely messed with ultra-linear amps, but I have measured the output power of some and yes it does increase the power output before it distorts a lot.

it shifts the grid line between what it would be for a pentode and a triode configuration, the way I understand it is that it makes the grid lines be more perpendicular to the load line so it will be able to swing more before cutoff and grid clipping.
or simply, a triode would distort when the screen voltage getting lower reduces the transconductance of the tube.
while a pentode would distort when the screen voltage was way higher than the plate resulting in large screen current. which will also keep the plate from getting down lower in voltage (since the screen instead atracts the electrons more)
so an ultra linear design is a fine balance where the screen gets pulled down just enough so it doesn't start pulling tons of current, and that it's not pulled down so low as to start really clipping the signal.

there's amps that used these, for example some fender silverface twin reverbs being rated at 135watts with 4 6L6GC's or fender super twin reverb, rated at 180watts using 6 6L6GC's, also using a cathode follower to drive the output tubes

and any output design can suffer from control grid clipping, that is when the control grid is at 0V and starts to draw current (starting to act as a plate, attracting electrons from the cathode), high output impedance of a phase inverter will get clipped by this current. but using a cathode follower which is a buffer, that has a low output impedance will be able to overcome this grid current and push the grid voltage higher, and make the tube conduct more. just that at a certain point the grid melts

 

[Dirty-D]

if your goal is to make it sound more like a 100 watt, l don't think you can. 100 watters have the Dagnall PT and the 50 watters have the Drake PT and they just sound different .lt's true ( at least to my ears). I suggest go get a 100 watter. l think they sound better anyways

 

[neikeel]

if your goal is to make it sound more like a 100 watt, l don't think you can. 100 watters have the Dagnall PT and the 50 watters have the Drake PT and they just sound different .lt's true ( at least to my ears). I suggest go get a 100 watter. l think they sound better anyways

You can get both 50 and 100 with both Drake and Dagnall depending on year of manufacture so that does not hold up.
I agree the 100 has more grunt.
The OP is says he is aiming for more grunt without $$.
Cheapest way is KT88s and some voltage tweaks, IMO 430v PT vs 450v PT is an expensive route, depends on if you really know what you want and need (I don't think he really knows).

 

[william vogel]

that's not really how tube life works, if you run the tube hard the cathode will start stripping it's coating off and destroy the tube quicker, if not it will still slowly evaporate.
so a tube that has twice the max plate dissipation means nothing in tube life, since the plate isn't what ages in the tube, it's the cathode.
for example an EL500 tube has a huge cathode to allow for a lot of current flow across the tube, but the plate dissipation is only like 18.5w iirc, it was designed for use in horizontal output circuits in TV's where the tube only needs to supply a short pulse of current instead of continuously conducting like in an audio amp.
if you used such a tube in an audio amp, you could even red plate the tube and it will last a long time (unless the plate melts) since the cathode was designed for a ton of emission.
however if you used like an EL34 which has a considerably smaller cathode in a horizontal output circuit, the cathode would just be nuked.

what I mean is that a larger cathode will last longer, a bigger plate tho wont.
higher plate dissipation demands a bigger cathode so it lasts at all, but the cathode wont always be matched to the max plate dissipation.

that's why for example the EL500 will red plate so hard that it melts the glass envelope; the cathode isn't limiting the current flow, nor is it stripping super fast.
while it's much less common on tubes like the EL34; cathode wont be able to emit any more electrons and will limit the current, and then before the glass envelope melts the cathode will already have stripped the coating off so much that emission will drop so much that the plate starts to cool down.

long lasting tubes would be for example:
6550; EL500; 6CB5A; 6LF6; 6LQ6; or many other horizontal output tubes.
if you look at Output-Transformer-Less amp designs, they will use tubes with huge cathodes so they can get a lower output impeadance, tubes like that will usually last way longer than normal audio amp tubes.

My relative remarks about the lifespan of the KT88’s vs the EL34’s had nothing to do with plate dissipation. It was about cathode current. The KT88’s can pass more current, and handle more voltage. We aren’t comparing Mullard XF2 and Genelex KT88’s, we’re comparing current production tubes. I can be wrong and that’s exactly why I explained that I didn’t have any real data to prove my claim. But a pair of KT88’s can run circles around EL34’s. KT88’s can produce 100 watts class AB to EL34’s 50-60 watts class AB. I love EL34’s, my favorite tube but KT88’s are just a stronger workhorse.
By the way, your wrong about ultra linear. Ultra linear never produces more power than pentode operation in true apples to apples comparison with equal voltage and impedance. Ultra linear is lower distortion but because it’s a split between pentode and triode operation, it produces less power.

 

[ezsemanate]

don't the super leads have 47k on the 8ohm tap for NFB?

Early ones yes. Punchy 73+ specs 100k 4 ohm. At least the 6550 USA amps Ive owned.

 

[Pete Farrington]

so a tube that has twice the max plate dissipation means nothing in tube life, since the plate isn't what ages in the tube, it's the cathode.

Note that GEC info directly correlates valve life to anode dissipation, see under ‘life performance’ p3 of https://tubedata.altanatubes.com.br/sheets/084/k/KT66_GEC.pdf

there's amps that used these, for example some fender silverface twin reverbs being rated at 135watts with 4 6L6GC's

Note the higher, stiffer HT supply. That’s where the uprated power output comes from. I think that the 12.5% g2 taps (my measurement) off the OT primary are coincidental to that. I think it’s incorrect to describe 12.5% taps as UL, I don’t think Fender described them as such.

fender super twin reverb, rated at 180watts using 6 6L6GC's, also using a cathode follower to drive the output tubes

Note the 33k grid leaks of the output valves. They put a heavy load on the LTP phase splitter outputs. CF buffers are probably necessary to drive sufficient signal level across them.

 

[sdn25]

Cheapest way is KT88s and some voltage tweaks, IMO 430v PT vs 450v PT is an expensive route.

What voltage tweaks are you talking about here? Yeah the PT route is expensive, if I can avoid it that would be great.

 

[sdn25]

@neikeel @Pete Farrington @william vogel I may have found a super super cheap solution. I remember once I had the amp on the 230v tap with 240 from the wall and I remember measuring 477-480 on the plates. However, the heaters then ran at 6.7V I just wanna know whether 6.7v heaters are detrimental. This is the easiest solution if it’s safe since it costs $0

 

[william vogel]

@neikeel @Pete Farrington @william vogel I may have found a super super cheap solution. I remember once I had the amp on the 230v tap with 240 from the wall and I remember measuring 477-480 on the plates. However, the heaters then ran at 6.7V I just wanna know whether 6.7v heaters are detrimental. This is the easiest solution if it’s safe since it costs $0

Run it like that for a session and see what you think. You need to adjust the bias so it’s not set too hot. Set it at 17 watts dissipation and make sure you deduct screen current. It’s only 6.3% elevated heater voltage. That’s within the +/- 10% tolerance.

 

[sdn25]

and make sure you deduct screen current.

What do you mean by that

 

[sdn25]

On the 230V tap it only reads 450V on the plates. Now I recall the 477-480V was on my old OT. The heaters also don't read 6.7V im not sure why. They only read 6.36

 

[william vogel]

What do you mean by that

If you use a device to measure cathode current for bias, either 1 ohm resistors on the tube sockets or bias probes or whatever. If it’s measuring cathode current, you must deduct the screen current from the measurement to get true plate current. If you measure plate current by knowing the DCR of the output transformer primary and you measure voltage drop across the output transformer, that’s true plate current but if it’s from the cathode, the cathode current is the sum of both the plate and screen so you need to deduct the screen current. It’s typically about 5mA so at 480 volts and 17 watts dissipation you need the bias current of 35mA. If you measure 35mA at the cathode that’s really just 14.4 watts if the screen is drawing 5mA.

 

[william vogel]

On the 230V tap it only reads 450V on the plates. Now I recall the 477-480V was on my old OT. The heaters also don't read 6.7V im not sure why. They only read 6.36

It’s probably because the mains voltage is really close to 230 volts right now.

 

[sdn25]

It’s probably because the mains voltage is really close to 230 volts right now.

I measured the mains its at 239 right now

 

[sdn25]

If you use a device to measure cathode current for bias, either 1 ohm resistors on the tube sockets or bias probes or whatever. If it’s measuring cathode current, you must deduct the screen current from the measurement to get true plate current. If you measure plate current by knowing the DCR of the output transformer primary and you measure voltage drop across the output transformer, that’s true plate current but if it’s from the cathode, the cathode current is the sum of both the plate and screen so you need to deduct the screen current. It’s typically about 5mA so at 480 volts and 17 watts dissipation you need the bias current of 35mA. If you measure 35mA at the cathode that’s really just 14.4 watts if the screen is drawing 5mA.

Okay so there are bias test points on the back of my amp which are connected to a 1ohm resistor to measure bias, Ive set it to where it reads 26mV at those test points using my DMM. With what you said it mind is the actual bias current only 21mV?

 

[Pete Farrington]

I just wanna know whether 6.7v heaters are detrimental

It’s not ideal, and the higher HT will increase the voltage and dissipation stress on the output valves.
But every new amp sold in the UK etc for the past 20 years should have been set up for 230V, though the actual mains is 240V.
That’s a lot of Hot Rod Deluxes etc. When I service them I reset them to 240V, to take some stress off everything.

But if it gets you where you want to be, then give it a try.

 

[ELS]

My relative remarks about the lifespan of the KT88’s vs the EL34’s had nothing to do with plate dissipation. It was about cathode current. The KT88’s can pass more current, and handle more voltage. We aren’t comparing Mullard XF2 and Genelex KT88’s, we’re comparing current production tubes. I can be wrong and that’s exactly why I explained that I didn’t have any real data to prove my claim. But a pair of KT88’s can run circles around EL34’s. KT88’s can produce 100 watts class AB to EL34’s 50-60 watts class AB. I love EL34’s, my favorite tube but KT88’s are just a stronger workhorse.
By the way, your wrong about ultra linear. Ultra linear never produces more power than pentode operation in true apples to apples comparison with equal voltage and impedance. Ultra linear is lower distortion but because it’s a split between pentode and triode operation, it produces less power.

I didnt know that power output of an amplifier was meassured at 100% THD. in that case just connect the speaker to mains and call that your power output.