New design amp problem with KT66's

[rtc]

Hi,
I've recently built a Root 66 clone which I didn't like (lifeless and dark sounding), so I decided to modify it , preserving the most expensive items.
I changed the preamp and PI to a Fender Super Reverb type without reverb, with a 12AX7 and a 12AT7, keeping just one treble, one bass and one volume control.
I also added a choke, and installed a new 20 mfd cap node. I enclose the schematic for those interested.
I biased the Valve Art KT66's at 36 mA, as plate voltage was about 460 V.
Plates voltages on 12AT7 were about 260Vdc and 250 Vdc on the 12AX7.
So yesterday I plugged in and liked very much the result. The amp was brighter and at 3 o'clock from the volume it started to crunch but in a very pleasant way (very blackface). I was shocked at how high was the volume at only 3, at 6 it was incredibly high, so I think there's a lot of signal on the power tubes. After playing for about 20 minutes without any problem I turned the volume up and at about 10, it made a funny noise, one of the KT66's went blue and died. I switched it on again and this KT66 was experimenting "fireworks" inside. It seems like arching.
Do you think it's just a tube failure (the tube was almost new)?
Do you think there's something wrong with the design?

Thanks for your help.

[VacuumVoodoo]

KT66 needs large gridstopper in few tens of kohms range and a well regulated/filtered screen grid supply some 50-70V below plate B+. Otherwise it will make a very powerful RF oscillator.
RDDS (Read the Damn Data Sheet)

[rtc]

Thanks,
Yes if you look at the data from TDSL it says no more than 500V plate, and no more than 400 V screen.
Original Route 66 has 1K5 grid stoppers and screens attached to UL taps (70%) through 470 ohm resistors.
JTM 45 KT66's amp has 5K6 grid stoppers and two 470 ohm resistors in series with one 1K resistor just after the choke.
So neither of these designs is very conservative but they works.
Could be the combination of 1K5 grid stoppers, 470 ohm screen resistors attached to choke (ala Fender), very aggresive, sufficent to ruin the tube?
But I think that I can keep the design and use 6L6GC, as it's almot identical to Blackface designs.
Or even use EL34's and only change bias.
Am I wrong?

[mhuss]

I don't know that the original GEC KT66s had this problem, but the modern ones definitely do. I recommend 5k6 minimum grid stoppers and 1k screen resistors. These are similar in spec to the 6L6GC, but have their own idiosyncrasies. I love these tubes, by the way!

--mark

[rtc]

YEs, I had the screens of the tubes attached to the UL taps of the Heyboer OT (I think 70%). When I installed a choke on it, as it hadn't, I forgot to change the screen resistors and left the older 470 ohms.
Yesterday I installed 1K-5W Dale resistors on them, changed the feedback resistor from 820 ohm to 2K7 (attached to 8ohm tap) and installed a pair of NOS Philips/ECG 6BG6GA (with appropiate converters), which in fact are the tubes I use in my tweed Bassman clone and, again, the sound was very good, but I have to improve the following:

-Gain is too high, and it starts distorting at 3.
-Bass is a somehow boomy, not tight.

I will try with different cathode bypass caps (in the rest of my amps I prefer values ranging from 1-3.3 mfd, Solen PP non-polarized type).
Perhaps I will change PI signal caps from 0.1 to 0.022.

Thanks for your opinions.

[Fischerman]

rtc,
A couple of things.
First, most BF Fender amps used a very small PI input cap and you have a .022uF there. That's really big for this circuit. Some BF amps have as low as 500pF there and many have .001uF. That could contribute to the boomy bass.

Look at a schematic for the Normal Channel of a BF Fender Super Reverb (or similar) and notice a few things:
1) There is a 220k series resistor after the coupling cap for the second stage.
2) There is another 220k resistor for the 'other' channel that connects to the 50k Tremolo Intensity pot which is grounded at the other end. That 50K pot is actually in parallel with the 100k plate load resistor for the second stage of the Vibrato channel (this is as far as signal is concerned since B+ is AC ground). So all that adds up to about 253k to ground.

So you effectively have about a 220k/253k voltage divider after the second stage coupling cap before it hits the PI. You're just going straight into the PI so you have quite a bit of extra gain there.

EDIT: Also note that your second stage doesn't have it's own coupling cap and the AC load on the second stage is just that 1M grid resistor for the PI input. But in the Fender schematic you'll see that the load on that second stage is closer to 220k+253k=473k. This probably doesn't make a huge difference but it's still a difference that's probably noticeable.

[Fischerman]

JTM 45 KT66's amp has 5K6 grid stoppers and two 470 ohm resistors in series with one 1K resistor just after the choke.
So neither of these designs is very conservative but they works.

Remember that since you have the current of both screens flowing through that 1K 'shared' resistor after the choke...that it has the same voltage drop as a 2K resistor placed in series with each screen. So if you have a 1K shared followed by a 470 for each one...the voltage drop to the screens would be the equivalent of a 2.47k resistor on each screen. So that arrangement is a bit more conservative than it looks.

[rtc]

Thanks for the info.
Yesterday I tried changing the first cathode bypass cap from 25mfd to 1mfd Solen PP, and modified the tonestack (0.1 to 0.022, and 0.022 to 0.047), using the Tonestack Calculator 1.3, and sound improved, as pots now work between 3 and 7 o'clock.
Today I will try modifying the PI input cap, and including the voltage divider.

[rtc]

I modified the signal cap afetr the second stage, installed a 220K/220K voltage divider, modified the power rail resistor to get the same voltage values as in a Blackface SR.
I've tried 0,01 and 0,001 input PI caps, as both values are used on different versions of SR and 0,001 works best. Makes the amp tighter.

I've tried 12AT7 and 12AX7's on PI, and 12AX7 sound harsh and distorts more. 12AT7 sounds better.
I like a lot the sound now, it breaks up early (I am using 4x10 Weber P10R's) and very bluesy. A good match for my Strats.
Byut i had the same problem again.
I turned up the volume full, and the fuse blowed. Replace it but then the amp didn't sound and blowed it again. After taking a closer look I found the origin, which had made the KT66 blow too:
The Heyboer OT I am using has 8K, 6K6 and UL primary taps. As I only use the 6K6 taps, the others were mounted on terminals. One of the 8K terminal tap had an small crack on the insulator and was arching to ground but only at high volumes.
The 6BG6 tubes are much stronger than KT66 and they kept working before the fuse blowed. I fixed it and now the amp is working OK.
Thanks

[stoo]

disregard

[billv]

Resurrecting this thread a lil' bit.

I'm looking at a Leslie 147a amp, with two 6550s. The screens are dropped 100v by an OC3.

(OC3: Basically the tube equivalent of a zener. This one's an RCA. Very cool purple glow!)

Dropping the screens ~50 to 70v was mentioned. Could I drop them 100v? Well - I know I could Would that affect tone, however?

[muchxs]

Resurrecting this thread a lil' bit.

I'm looking at a Leslie 147a amp, with two 6550s. The screens are dropped 100v by an OC3.

(OC3: Basically the tube equivalent of a zener. This one's an RCA. Very cool purple glow!)

Dropping the screens ~50 to 70v was mentioned. Could I drop them 100v? Well - I know I could Would that affect tone, however?

What are you trying to do? Run KT66s in a Leslie?

Dropping screen voltage drops output power. Yes it affects tone. Just dropping screens takes the amp from sharp, crisp and right there to "What's missing?".

The Leslie 147 is a very conservative circuit. It only makes around 35 watts rms. The typical organ player is looking for more not less, it seems guitar players can blow a Leslie right off the stage.

If you have a 147 amp keep it a 147 amp or do the few small mods to convert it to a 122.

[billv]

Just trying to understand that circuit and compare to a planned JTM45 build.

Was considering throwing an OC3 in there if tone wasn't compromised. I'm strangely attracted to that purple glow!

This particular 147 will stay as is, i.e. in the 145. I have a bare 46W/122/147 cabinet to fill, however. Later on down the line, I'll build up a custom 147, with the 122 mods as you suggested. Hopefully (unless the iron is different) I can make it switchable.

Thanks for the help. I'll keep that pesky OC3 out of the screen circuit!

[muchxs]

Just trying to understand that circuit and compare to a planned JTM45 build.

Was considering throwing an OC3 in there if tone wasn't compromised. I'm strangely attracted to that purple glow!

A 147 is a whole different animal compared to a JTM45. Different tubes, different power supply, different bias scheme (check it out!), different in almost every way.

If you really want to get a handle on your JTM45 look at the tweed Bassman. If you like strange glow go with the 83 mercury vapor rectifier tube in the early version of the 5F6.