PP valve amp KT88 BIAS-question

[cfortner]

Hi,

I am wondering about the calculation of Rg in a pp valve amp with KT88 and fixed BIAS.

Summing the resistors from G1 to ground results in the value of Rg (KT88 100.000 ohms max., KT120 51.000 ohms max.).

My question: Have I to include the value of the grid stopper? Or is Rg calculated without the grid stopper?

Thanks

Christian

[pdf64]

A grid stopper is a series resistive element in the path between grid and its cathode, so it should be included in the Rg1-k calculation.

[cfortner]

Thanks a lot, that doesn't make it easier to match with the data sheet values...

The most PP-amps I saw have much larger Rg, but are working fine. What happens with a PP KT88 with a Rg of 200k, f.e.?

[tubeswell]

Hi,

I am wondering about the calculation of Rg in a pp valve amp with KT88 and fixed BIAS.

Summing the resistors from G1 to ground results in the value of Rg (KT88 100.000 ohms max., KT120 51.000 ohms max.).

My question: Have I to include the value of the grid stopper? Or is Rg calculated without the grid stopper?

Thanks

Christian

Assuming the grid stopper is between the grid leak and the grid, then technically you should sum everything up. But a 4k7 grid stopper won’t add much to 100k, while 47k grid stopper would.


The main issue with 100k (or 51k) grid leak resistance is lossy impedance-bridging, i.e., with a signal going from a relatively high output impedance from the previous source  (driving stage) into the relatively low input impedance of the output tube grid. The source impedance forms a voltage divider with the load impedance. Where source impedance is higher and load is lower, there is more signal attenuation. This is why it’s better to drive KT88s and KT120s with a source impedance buffer (like a DC-coupled cathode follower, where the cathode resistor of the driving stage and the grid leak resistance for the output tube are one and the same).

[cfortner]

Thanks, that makes it clearer to me.

I'm planning a 500W PP with six KT120, fixed BIAS and plate voltage of 830-850V. It's a poweramp near to the valve limits and for this reason I want to build it with a Rg of below 51k per tube (17k over all per side).

Thats not so much and I plan an E88CC as cathode follower between the PI and the KT120.

But again, what happens when Rg is too high? You told me, lower gain, ok. And beside this? Instable BIAS? Other affections?

[tubeswell]

what happens when Rg is too high? You told me, lower gain, ok. And beside this? Instable BIAS? Other affections?

Build-up of excess charge on the grid leads to bias instability

[cfortner]

Thanks!

[pdf64]

… what happens when Rg is too high? You told me, lower gain, ok. …

You may have misunderstood. As the Rg value increases, the potential divider effect formed with the preceding stage will reduce, signal level at the grid will increase.

[AlNewman]

This is good info

[DummyLoad]

Thanks, that makes it clearer to me.

I'm planning a 500W PP with four KT120, fixed BIAS and plate voltage of 830-850V. It's a poweramp near to the valve limits and for this reason I want to build it with a Rg of below 51k per tube (25,5k over all per side).

Thats not so much and I plan an E88CC as cathode follower between the PI and the KT120.

But again, what happens when Rg is too high? You told me, lower gain, ok. And beside this? Instable BIAS? Other affections?

4 or 4 pairs? 4 (two pair) won't yield 500W.


--Pete

[cfortner]

Sorry, six KT120 (3 pairs in PP), resulting Rg over all 17k, I corrected this in my earlier post.

[PRR]

Do you have experience in >100W amps? At that level, stuff goes wrong very fast.

[cfortner]

Hi,

yes, I built a Hiwatt DR405-clone with 6x 6550 and real 300W from scratch, repaired and modified a Mywatt 400 (german DR405-clone) up to real 400W with 6x KT88 and Va 800V, Vg2 410V (both amps are working fine).

I tried the KT120 in the Mywatt 400 with Va 800V Vg2 410V and got nearly 500W. But the BIAS-circuit, which determines Rg, is too hi z (Rg more than 130k caused by trimmpots with 50k). At high level performance near the limits it seems to be little instable, BIAS begins to float.

Changing the complete BIAS-section seems very complicated in the Mywatt circuit, I'm satisfied with real 400W from six KT88, which is more stable due to the recommended Rg of 100k.

But the KT120 are fascinating and I want to build an amp 6x KT120, Va 850V, Vg2 450-500V, toroid 1000VA and 400W OT Raa 1,45k. For this amp I need to bring Rg down to 51k for each KT120 = ca. 17k for three KT120 in parallel.

This is possible with a 10k trimmpot-circuit for every KT120, which delivers a BIAS-range of -36 to -60V.

The preamp will be a Hiwatt with the original LTPI and a DC coupled cathode follower with a 12BH7 or ECC99. The output impedance and current ability of these valves are strong enough to drive the six bottles, separate BIAS possible.

[sluckey]

When you edit a lengthy post it would be nice to indicate why you edited it. Could be as simple as adding a comma, or a capital letter, or it could be as involved as editing the content that might change the meaning of the post. It's a little disappointing to re-read a long post and not be able to tell if anything important has been changed. I'm more inclined to just turn the page.

[pdf64]

… I want to build an amp 6x KT120, … For this amp I need to bring Rg down to 51k for each KT120 = ca. 17k for three KT120 in parallel.

This is possible with a 10k trimmpot-circuit for every KT120
…

If each output valve has its own bias trimmer, it seems to me that each must also have its own (50k) grid leak resistor and coupling cap?

Yes the AC loading on the driver will be 17k.

[cfortner]

Sorry, I corrected the post two or three minutes later and did not assume everybody just read it in this moment. Will not happen again.

Rg is summing to ground this way:

grid stopper 1k
injection resistor 33k
trimmpot max 10k, in the middle 5k
resistor from trim to ground 8.2k

All together something between 45 and 50k. Means a load of 15-17k for 3 KT120 parallel.

Edit: corrected two wrong written words

[pdf64]

By ‘injection resistor’ you mean grid leak?

[DummyLoad]

For that design, you should seriously consider a Solid State driver,  (source follower) with a grid leak load in the teens, e.g. around 12-15kohms.


--Pete



 

[cfortner]

Yes, grid leak resistor. I‘ll post a schematic.

I want to avoid a SS driver, may be difficult.

[tubeswell]

borrow a proven design e.g. see attached

[cfortner]

I built an exact clone of this amp, Hiwatt DR405. Works fine with 6 KT88 or 6550.

Using KT120 is more difficult because of matching problems. Better to bias every KT120 seperate.

I drawed a schematic with the Hiwatt section after the master volume with LTPI and dc-coupled CF, but ac-coupling CF - KT120.

The BIAS-circuit, which determines Rg, creates a Rg of ca 42k, correct for KT120, which is assuming not more than 51k.

Three KT120 in this circuit parallel means 16-17k load for the PI. Heavy.

I have to look, on which point the 12BH7 delivers around 130V pp into the load of 15k. Clean.

[PRR]

cfortner, you seem to have posted the 150x90 thumbnail. Is there a bigger version we can read?

[cfortner]

Sorry! Better?