c3m + 300b + miller capacitance

[Klimon]

Hi,

I'm considering to use c3m pentode as driver for a 7236 dual triode se, the latter needs 100Vpp swing and the former can deliver 115Vpp

BUT

When calculating the low pass -3db point due to miller capacitance I get around 19khz = not good.

What baffles me is that the c3m gets used by several people to drive a 300b; relevant c3m variables won't differ much in those circuits and in mine (260V B+/10K load/ ca. 200k Ri) AND the 300b has even worse Cag than 7236 (15 vs. 10)... So... Could anyone shed a light on this enigma? Is c3m-300b a bad miller capacitance mariage or should I practice a little harder on my miller calculations (probably!)?

Thanks alot!!!

Simon

 

[JoshK]

Excellent question. Very relevant to a project I have going. Woud also love to hear some insight.

 

[Klimon]

http://www.diyparadiso.com/triple1.htm

Triple 300B, c3m driver = 3* miller capacitance

Still chewing on the same teaser

Simon

 

[pmillett]

Well... on Benny's triple 300B design, if I calculated correctly, I get a f(-3dB) of around 23kHz.

You have to include the plate load resistor of the pentode (none, in Benny's design, as it has a plate choke) in parallel with the grid-ground resistor(s)s on the 300B(s) (a total of 27.5k in this case) in the output impedance of the driver. This dominates the Zo of the driver, but it is also in parallel with the Rp of the pentode.

On the single 300B, freq response depends on what those resistors are. With a typical resistive load on a c3m there should be no problem...

Pete

 

[Klimon]

thanks a lot Pete, I put the plate load and Rp in series in my calculations... c3m looks perfect for the job

Simon

 

[kevinkr]

Did you put the load load resistor and the tube rp in series or parallel? The C3M source impedance is actually the thevenin equivalent of rp and the plate load resistor in parallel, not series. This incidentally would improve the picture still further.

 

[Klimon]

Did you put the load load resistor and the tube rp in series or parallel?

Yep, that's the mistake. I'm not exactly following you with mr. Thevenin but common mathematical sense tells me that my error should lead to a big difference in end result......... and then there's the comfort of relying on predecessors doing similar things... If the wheel rolls smoothly no need to question it's roundness

Simon

 

[triode_al]

As to a pentode driver to a 300B, I tried 6SJ7 connected to a 300B directly with a small capacitor of 100 nF PIO: bandwidth 15 KHz with Ra of 67K.
That's a lot of Miller. It surprised me, the 300B having the low mu and I had a 2K1 transformer.
By the way, a 1 uF PIO made this even worse.
I had to insert a ECC82 buffer and now it is 150 KHz; and it can swing 180 V pp.

 

[disco]

As to a pentode driver to a 300B, I tried 6SJ7 connected to a 300B directly with a small capacitor of 100 nF PIO: bandwidth 15 KHz with Ra of 67K.
That's a lot of Miller. It surprised me, the 300B having the low mu and I had a 2K1 transformer.
By the way, a 1 uF PIO made this even worse.
I had to insert a ECC82 buffer and now it is 150 KHz; and it can swing 180 V pp.

Tried a lower value for grid leak resistor? Tried reversing OT polarity?