12AT7 Inverter in Poseidon Mark IIIs

for Dynaco Mark II/III/IV and DIY PP monoblocks

Postby WLK » Mon Dec 29, 2008 8:07 pm

As per the last post I have changed the two plate resistors R10 & R11 to 56K @ 2 watt resistors, reinstalled the 12AT7 as the phase inverter in the Poseidon circuit, powered up the amps using a Variac and obtained the following results:

B+ 485 VDC, C+ 345 VDC, D+ 310 VDC

Current set resistor is 10 ohms resulting in a current of 3.5 MA per plate at 35* C.

12AT7 - Pin #1 = 245 VDC Pin #6 = 260 VDC Pin #3 & #8 = 3.9 VDC

Amps run stable up to maximum power.

Amps sound great - tight bass, well defined mids and highs.

I then changed the current set resistor to 15 ohms resulting in a current of 2.3 amps per plate at 35* C, slowly powered up the amps using a Variac and obtained the following results:

12AT7 - Pin #1 = 310 VDC Pin #6 = 340 VDC Pin #3 & #8 = 5.9 VDC

Shut down amps because the plate voltages were above the maximum specified for a 12AT7.

Shannon, why does the voltage at Pin #3 & #8 have to be 9 to 13 VDC?

At 3.9 VDC at Pin #3 & #8 the amps run stable to full power and sound very good.

Should I increase the plate resistors back up around 100K, use the 15 ohm cureent set resistor to set the current at 2.3 ohms per plate and try the test again?

Any other suggestions?

Regards
WLK
WLK
 
Posts: 8
Joined: Sun May 13, 2007 5:19 pm
Location: Vancouver, BC Canada

Postby EWBrown » Mon Dec 29, 2008 10:08 pm

THe cathode voltage at 3 & 8 only has to be around 9 to 13VDC with 12AU7s or 6CG7s, etc, which are lower mu tubes.

I've run 6EM7 VAs with the same CCS circuit, using the LM334Z, which set the conjoined cathodes around 2.6V above ground.

It works and sounds OK, but it cuts really close to clipping just as maximum output power occurs. As long as the peak audio signal which travels between the cathodes is llower than the DC offset, there will be no clipping or distortion issues.

Low mu LTPI tubes will set the cathodes at a higher voltage, as compared to high-mu tubes, it's a result of the tube's transconductance, and plate resistance (and the resulting mu or amplification factor).

If it sounds good at 3.9V, then go for it, there is really nothing wrong with iunning it at that voltage..

I see that pin 6 runs at 260VDC, about 15 VDC higher than pin 1. This is normal for LTPI circuits, as there is some normal imbalance in them. This can be corrected by increasing the plate resistor connected to pin 6 by around 4.3K ohms, either connect a 4.3K resistor in series,, or change the 56K resistor to 60 to 62K (the closest standard 5% value). That will lower pin 6 to around 240 to 245V, which is closer to the other plate's voltage.


12AT7s have plenty of plate dissipation, so 3.5 mA at your normal plate voltages is well within the plate dissipation ratings, about 875 mW, compared to 2.5W.

There is another "work around" for this, but it starts getting a bit complicated, and that is to use the same CCS and 10 ohms curernt setting resistor, but then reference it to a negative DC voltage, rather than 0V ground, something around -6VDC should work nicely.

That will offer lots of extra voltage "headroom". The cathode currents will still be the same, but the cathodes will appear to be sitting at 9.3 VDC above the new "virtual ground" reference. The same approach was used in earlier LTPI designs, the "tail" resistor would be between 12K and 18K, and referenced to a negative voltage somewhere between -50 and -100 VDC.

/ed B
Real Radios Glow in the Dark
User avatar
EWBrown
Insulator & Iron Magnate
 
Posts: 6389
Joined: Wed Mar 19, 2003 6:03 am
Location: Now located in Clay County, NC !

Postby WLK » Tue Dec 30, 2008 12:17 am

Ed

Once again thanks for the excellent information.

You have answered all my questions very clearly and I now understand the logic of the Poseidon phase inverter.

As you suggested I will increase the value of the plate resistor connected to pin #6 to 62K to help balance the plate voltage.

I have been listening to the amps with various types of well recorded music for the past six hours and I am extremely please with the results. As previously mentioned the bass is strong and very well defined (no old school flabby tube thump) and the midrange and highs are balanced and also very well defined.

Thanks for all the valuable time that you have spent helping me successfully complete this project.

Best Wishes
Wayne
WLK
 
Posts: 8
Joined: Sun May 13, 2007 5:19 pm
Location: Vancouver, BC Canada

Previous

Return to poseidon

Who is online

Users browsing this forum: No registered users and 14 guests