Modification for top plate reading idea

[Paully]

Just floating an idea for modifying my Paramounts with the new the driver board.  Was thinking about drilling a hole for a tube socket to be able to do do voltage checks on the driver and output tube (kind of like a Dynaco ST-70 bias set-up).  Then was also thinking about putting a pot such as a clarostat 10-turn mounted on the plate to replace the small pot on the driver board.  Any pitfalls or issues that I might be missing that would make this difficult to pull off?  Any suggestions?

 

[Paul Joppa]

The high voltages are inaccessible from the outside for obvious safety reasons.

Connector bA on the board goes to ground with a black jumper. If you replace that with a 3.3K resistor, you can read the voltage at bA which will be 1/100 of the high voltage - thus much safer to bring it outside. Similarly, a 3K resistor in series with the 300K OB to -reg (B side) will give 1/100 of the driver plate voltage. Both can be protected with 5v to 10v Zeners in case of an internal fault; the Zeners will be out of the circuit in normal operation.

Use shielded twisted pair wiring on the trimmer to protect it from noise pickup. The potentiometer case (if metal) and the shield can be grounded to the chassis at the mounting point. This is a low voltage (the driver cathode bias) and should be safe.

Let us know if you do it and if it works!

 

[Grainger49]

Paul,

A couple of questions about the information below since I will be working with Paully on this.  I also don't have access to the upgrade instructions and voltages.

What wattage rating should the 3.3k resistor be?  Since you suggest a 5V or 10V Zener I expect less than 5V there.  The Zener should be reverse biased so it doesn't conduct unless the resistor opens, right?

What wattage rating should the 3k resistor be?  I would guess it is not going to be grounded so we need to bring wiring across it out to measure, right?  Both voltages will be brought out to a 7 pin tube socket that has enough clearance to insert his meter leads into.

I think Paully has TSP in his spare parts, the cases are plastic (Clarostat)and we will ground one end of the TSP.

Thanks for all the help on this one.  One day I will bring my Paramour hum pots out on the top plate.

I will take pictures of this endeavor.

  .  .  .  Connector bA on the board goes to ground with a black jumper. If you replace that with a 3.3K resistor, you can read the voltage at bA which will be 1/100 of the high voltage - thus much safer to bring it outside. Similarly, a 3K resistor in series with the 300K OB to -reg (B side) will give 1/100 of the driver plate voltage. Both can be protected with 5v to 10v Zeners in case of an internal fault; the Zeners will be out of the circuit in normal operation.

Use shielded twisted pair wiring on the trimmer to protect it from noise pickup. The potentiometer case (if metal) and the shield can be grounded to the chassis at the mounting point. This is a low voltage (the driver cathode bias) and should be safe.

Let us know if you do it and if it works!

 

[Paul Joppa]

There is a 330K C4S bias resistor on the board, and the 3.3K is in series with it, one end grounded. So if there is 450 volts at the power supply, then there will be 4.5 volts across the 3.3K resistor. Running 1.5mA the power rating should be larger than 0.00675 watts.  :^)

The 300K is grounded, and goes from the driver plate to ground. Sticking a 3K in the ground end gives 2 volts if the driver plate is at 200v. Again, no power issues. The Zener is just there to make sure that whatever connection goes outside the chassis, stays at a low voltage - yes, cathode to positive, anode to ground.

The bias adjustment on the driver occurs in the cathode circuit, where voltages are normally no more than 4 volts. The 431 will reject noise on that line, but it IS in the cathode, a high-gain position, so I suggested the shielded wire out of caution.

Get Paully to bring the manual; a glance at the circuit should make this more clear.

 

[Grainger49]

Thanks, I follow you completely.  I'll select some resistors from my backlog of resistors.  I'm doing the driving so the manual will be there.

 

[Paully]

Yep, I come up with the modification and I get the electrical engineer to drive here! 

 

[Grainger49]

Ok, looking at the schematic of the soft start installation I see that adding a 3k resistor would give 1/100 of the driver plate voltage.  If I read it correctly, a 3.3k resistor attached from bA to ground would give 1/100 of the voltage feeding the regulator half of the tube, + reg (B) and time delay and C4S feeding the driver plate.  

Does that voltage at Ba come into play when adjusting the voltage of of the driver plate?  That is, do we need to know this voltage?

There is a 330K C4S bias resistor on the board, and the 3.3K is in series with it, one end grounded. So if there is 450 volts at the power supply, then there will be 4.5 volts across the 3.3K resistor. Running 1.5mA the power rating should be larger than 0.00675 watts.  :^)

The 300K is grounded, and goes from the driver plate to ground. Sticking a 3K in the ground end gives 2 volts if the driver plate is at 200v. Again, no power issues. The Zener is just there to make sure that whatever connection goes outside the chassis, stays at a low voltage - yes, cathode to positive, anode to ground.

The bias adjustment on the driver occurs in the cathode circuit, where voltages are normally no more than 4 volts. The 431 will reject noise on that line, but it IS in the cathode, a high-gain position, so I suggested the shielded wire out of caution.

Get Paully to bring the manual; a glance at the circuit should make this more clear.

 

[Paul Joppa]

"... do we need to know this voltage?"

Yes, the driver plate voltage is adjusted to 1/3 of the power supply voltage as described in the upgrade manual (and the v1.1 Paramount manual).

 

[Paully]

It looks to me like we have a good understanding now of what it is we are up to.  We only have one last question left.  By doing this we will have access to two measurements, the driver voltage at a3 and the plate voltage at a2 (reduced by the resistors).  But we won't have the cathode voltage on top.  When the driver voltage is changed it also changes the cathode and plate voltage if I understand everything.  If we make sure we know all of the measurements the first time we set the driver voltage will measuring the two points in the future be enough to be confident that everything is in spec even if I change tubes?  I would think that the simple answer is yes (and that would be enough for me).  Thanks for all of the help Paul, we really appreciate it.  Looking forward to the upgrade (and the Magnequest iron!).

Paul

 

[Paul Joppa]

...By doing this we will have access to two measurements, the driver voltage at a3 and the plate voltage at a2 (reduced by the resistors). ... will measuring the two points in the future be enough to be confident that everything is in spec even if I change tubes?...

Yes.

Assuming the 2A3 meets its specs, the cathode voltage and current can be calculated from the grid and plate voltages. I did a bunch of spreadsheet simulations, whose object was to optimize the 2A3 operating point to match a 3K load impedance. This led to the simple rule that the grid voltage for a 2A3 Paramount should be 1/3 of the B+ voltage (which is almost the same as the 2A3 plate voltage). This rule gives a close enough result (within a few percent) over a range of +/-10 percent in power line voltage.

Due to 2A3 tube variations, these values cannot be much more precise unless you measure the mu and transconductance of the tube as well, and run a spreadsheet to calculate the optimum driver plate voltage. Mu and transconductance would have to be measured at the chosen operating point of course, and they will change somewhat as the tube ages, so (IMHO) great precision would be quite impractical!

Incidentally, if you use a 1.1K resistor instead of a 3.3k resistor attached from bA to ground, you will get 1/300th of the B+ voltage - so the two voltages should be made to match. Then you don't have to divide by three.

I should mention that he 330K power resistor on the board is a 5% part; you may want to measure it and make its small resistor is actually 1/100 (or 1/300) of the true value. I think this could be a small but useful improvement in overall accuracy.

 

[Paully]

We are doing it for a 45, but obviously the principle is the same.  It sounds like it will work so I think we will go ahead and have some fun with it.  Thanks for all the detail!