Home Tubes Asylum

Questions about tubes and gear that glows. FAQ

Review: Poindexter PP 6V6 Amplifier.....

207.89.249.52

I came across a pair of monoblock Hammond Jensen model G amplifiers manufactured in 1946. These amplifiers had each two 6J5GTs, four 6V6GT and a 5U4 rectifier. The 6J5 preamp was capacitor coupled to a 6J5GT phase splitter.

Knowing more elegant amplifiers existed, I reviewed a few different designs. One amplifier that was impressive was the Poindexter 6V6. A 5965 input phase splitter drove a pair of PP 6V6GT triode connected. The simple circuit has only two coupling capacitors and just a phase splitter/driver. No other preamp tubes were required. I do believe in simplicity or less is more at times. Performing research regarding the 5965 tube, I found the 5965 is a high quality tube for phase splitting service.

The difference in-between the Poindexter design and my modified design was the Hammond Jensen model G amps have four 6V6GTs. An interesting circuit in the model G amps was a pair of air wound small inductors that was placed in-between the 6V6GT anodes. I figured these inductors reduced some interaction in parallel push-pull tubes. Perhaps that parallel tube slight sonic smearing some report was addressed in this amplifier by the design engineer.

The model G amps have a LCLC power supply already in place with oil capacitors. The chokes were about 5HY with a DCR of 30 ohms. This amplifier platform is of high quality. I modified the power supply with a 47uf Sprague Atom after the first inductor. Then after the second inductor, I simply paralled the existing oil cap for 15uf of capacitance. At this point, the second capacitor is the B+ supply for the audio transformer. From this point, I installed a small 10HY choke with a DCR of 500 ohms. (The Poindexter design has a 1K resistor at this point). At the output of the 10HY choke, I installed an 8uf Sprague Atom capacitor for the 5965 tube B+ supply.

The Poindexter design required fixed bias for the 6V6GT tubes & -140 volts for the 5965 cathode. I installed two 120 VAC to 30 volt AC transformers of 10VA each and operated one transformer backwards from the 30 volt secondary to achieve a 120 vac source. Using a pair of Sprague Atom 40uf rated at 250 VDC, I installed a RCRC network. The first resistor 350 ohms was to smooth out the higher peak to average current in a capacitor input design. The second 350 ohm resistor feeding the second capacitor was to reduce AC ripple and hopefully target the -140 volts requirement. The 6V6GT grid bias used a full wave rectifier exactly the same as the -140 volt DC supply. I used a CRCRC circuit to keep AC ripple low with 2K, 50uf, 2K, 50uf, 5K pot and 5K resistor to ground. The last 50UF capacitor was installed on the pot wiper. I ended up with a -20 to -26 volt variable output.

To my suprise, I was expecting a 300 to 315 B+ supply. I measured 350 volts with a simulated load & 5U4GB rectifier. I installed a pair of 250 ohm resistors in-between the transformer high voltage winding and the 5U4 plates. The voltage dropped to 335 volts. Usually, 6V6GT tubes are limited to the 315 volts B+ as shown in the Poindexter design. I figured the 100K grid resistors was the upper limit for parallel fixed bias 6V6GT tubes & the extra grid capacitance of two tubes vs one would tax the 5965 output. So, a few extra volts of B+ was to my advantage or at least I figure such as long as the 6V6 tubes hold up.

Finally, I had a pair of MQ-465 (I believe) Dynaco ST70 type replacement audio tranformers in the parts bin. The primary is 4300 ohms impedance anode to anode. The secondary is 16/8/4 ohms. Using modified ProAc 2.5 Response speakers, I believe the impedance is closer to 6-ohms, not 8-ohms. I found the best sonics was using the 4-ohm taps. Thus, for 8-ohm speakers, a primary of 5500 to 6000 ohm is best. Poindexter also recommended this higher impedance. Keep in mind, poindexter has spent a lot of time tuning his design. I would carefully follow the recommended high quality parts as shown in the design.

Fortunately, I have a lot of tubes. Tube rolling is important in this amplifier like other amplifiers. The best input tube is an Amperex 7062 'pinch waist'that is a high quality replacement for 5965. Next, the Apmerex PQ 7062 is also a good tube with somewhat less tight bass response. An ok tube is an RCA 5965 triple mica if the Amperex are not available. I tried Sylvania metal 6V6 (awful), GE coin base 6V6GTA and Sylvania JAN CHS 7C5 from 1944. Yes, the 7C5 is a 6V6GT in a loctal socket. I give a very slight edge to the 1944 Sylvania, however the GE coin base are excellent tubes. If I had a lot of spares I would have stayed with the GE 6V6GT coin base. After trying some RCA 5U4GB, RCA 5U4G, RCA 5U4GB 'tall glass', Tung-Sol 5U4G and various 1950s to 1960s Sylvania 5U4GB, the Sylvania was best. The best Sylvania has the full size top mica & small bottom mica. As for that 335 volts, the 6V6GT sounds best at -23.5 volts bias. The tubes lose some midrange magic at -24 to -25 volts DC bias.

So, how does the amp sound? Excellent!

I would not have figured a tetrode strapped tube into triode operation could sound this good. I had a few vintage amps including Grommes 24PG, Dynaco Stereo 35, Fisher monoblocks, etc. All those amps had AuriCap replacement capacitors, other upgrades and many tubes auditioned for best sonics. Fact is, the Poindexter 6V6 is superior by a fair amount. The only amp I own that plays in the same league is a DIY PPP 6B4G triodes design. The 6B4G splitter & driver is more complicated than the Poindexter design, thus perhaps that 'less is more' handily applies in the Poindexter 6V6.


This post is made possible by the generous support of people like you and our sponsors:
  Schiit Audio  


Topic - Review: Poindexter PP 6V6 Amplifier..... - Jimmy 20:21:49 01/22/05 (7)


You can not post to an archived thread.