6P43P Strange Plate Graphs - Page 2 - diyAudio
Go Back   Home > Forums > Amplifiers > Tubes / Valves

Tubes / Valves All about our sweet vacuum tubes :) Threads about Musical Instrument Amps of all kinds should be in the Instruments & Amps forum

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 19th April 2011, 08:25 PM   #11
TheGimp is offline TheGimp  United States
diyAudio Member
 
TheGimp's Avatar
 
Join Date: Aug 2009
Location: Johnson City, TN
It is a 10bit D/A, in a PIC16F876.
  Reply With Quote
Old 20th April 2011, 04:37 PM   #12
TheGimp is offline TheGimp  United States
diyAudio Member
 
TheGimp's Avatar
 
Join Date: Aug 2009
Location: Johnson City, TN
Another set of plots from last night (my computer at home won't boot so I have to wait till now to post).

I installed a transformer for g2 with the return taken from the cathode instead of ground so the screen current is no longer going through the cathode sense resistor.

The initial plots look much better but still off a good bit from the data sheet. The knee is still wrong. I had previously wondered if the PIC can keep up with the dv/dt of the ac waveform used for anode drive. In retrospect I would expect different distortion if that were the case.

With 60hz, 400Vpeak, the dv/dt is 6.6V/us at zero crossing. If the sample rate were too slow around zero cross, I would expect compression of the waveform with a steeper onset to the knee and that is not what is happening.

In the zip is one file that is really screwy. The 6P15P just went weird with alternating plots no matter what I did with the G2 voltage.

Right now the G2 transformer is on the bench. I need to find one with a higher output voltage to mount inside.

Next up will be changing the cathode resistor from 10R to 1R to minimize cathode feedback, and increase the gain of the op amp ckt to compensate.
Attached Images
File Type: jpg 6P43P-E_2.JPG (161.7 KB, 157 views)
Attached Files
File Type: zip Tube Plots_2.zip (226.9 KB, 14 views)

Last edited by TheGimp; 20th April 2011 at 04:39 PM.
  Reply With Quote
Old 20th April 2011, 05:15 PM   #13
diyAudio Member
 
smoking-amp's Avatar
 
Join Date: Dec 2001
Location: Hickory, NC
You could try attenuating the plate voltage further to see if dV/dt is a problem. Is there a sample and hold before the comparator? I would think that a tracking D/A and comparator would have some accuracy problems with fast slewing waveforms without a sample and hold. I think Analog Devices makes such a S/H chip. If the digitized waveform is not stable, then I would check for some HF oscillation of the tube. If so, some grid stoppers and ferrite beads are in order.

It appears that the distortion is still related to screen current in some way. The screen supply needs to have low capacitance to ground since it is floating on the cathode current sense resistor.
__________________
Without MS, how would we live without all those hackers?

Last edited by smoking-amp; 20th April 2011 at 05:19 PM.
  Reply With Quote
Old 20th April 2011, 05:24 PM   #14
Yvesm is offline Yvesm  France
diyAudio Member
 
Yvesm's Avatar
 
Join Date: Jul 2004
Location: Ardeche
Mmmmh . . .
In my first attempt I envisged to use a PIC for A to D conversion but I was unable to sample current and voltage simultaneously (well, almost ...) even using the sample and hold properties and counting machine cycles.
This may explain that ?
Then, I went to a stereo sound board . . .

6P15 has no internal connection btwn K and G3 I know, I've paid
  Reply With Quote
Old 20th April 2011, 05:43 PM   #15
diyAudio Member
 
smoking-amp's Avatar
 
Join Date: Dec 2001
Location: Hickory, NC
The software algorithm used to do the A/D conversion can in fast slew cases be speeded up (provided there is a sample and hold feature) by using the logarithmic successive approximation technique. This homes in on the solution from most significant bit down to least sig. bit, bitwise. But requires 10 iterations thru the loop for 10 bits.

The other simple approach of using an up/down counter for the D/A is rather slower for random signals, but works quite well for smoothly slewing waveforms, up till the loop's max slew rate is reached. Only one iteration thru the software loop needed at the max slew rate of one count per sample. It is possible to combine the two approaches in a hybrid scheme for waveform tracking, that jumps the count ahead or back by 1/2/4/2/1 counts etc, ie, then back or forward 1 or 2 counts, then +/- 1 count to home in on the signal.

Even so, there really need be no issue with A/D speed for this application at all. You just need a memory table to integrate all the point measurements together over several sweeps (or use multiple pointers for the moving sample points on each stepped waveform). (a S/H is required though still ) Going this slow route can also enable averaging over multiple sample sets to remove noise, but too many data sets accumulated could lead to tube heating/drift problems.

UhOh...., I just realized that if there is only ONE S/H, then there is a timing issue between the voltage and current digitizations. One would need to add two outboard S/H units to synchronize them, or revert to the slow scanning scheme above with voltage digitizations and current digitizations on alternate sweeps, but this would require perfect identical software timing in each pass. The software would have to be designed to consume the same number of clock cycles on all software branches. Whew.. tough to code.
__________________
Without MS, how would we live without all those hackers?

Last edited by smoking-amp; 20th April 2011 at 06:12 PM.
  Reply With Quote
Old 20th April 2011, 05:55 PM   #16
TheGimp is offline TheGimp  United States
diyAudio Member
 
TheGimp's Avatar
 
Join Date: Aug 2009
Location: Johnson City, TN
At first I thought I had an oscillation problem with a 6P43P-e so I added ferrite beads on the grid, G2 and Plate at the tube socket. I also added a cap from grid to cathode per the manual on the tube test set.

The manual on the PIC shows a s/h with a tau of .42us maximum. The a/d converter takes 10 cycles to convert, so it is either ratiometric or SAR.

I just realized that since the input waveform is scaled before feeding the a/d, the dv/dt at the converter should be much less, probably around .00155v/us so compared to the s/h time constant the dv/dt of the anode voltage should not be an issue.

Back to scratching my head.

I may try attenuating the plate voltage any way just to see what happens. Plat supply is now seperate from all other power supplies so I can put a variac on it and dial it down.

Last edited by TheGimp; 20th April 2011 at 06:10 PM. Reason: a/d cycle requirement
  Reply With Quote
Old 20th April 2011, 06:05 PM   #17
TheGimp is offline TheGimp  United States
diyAudio Member
 
TheGimp's Avatar
 
Join Date: Aug 2009
Location: Johnson City, TN
Quote:
Originally Posted by Yvesm View Post
Mmmmh . . .
In my first attempt I envisged to use a PIC for A to D conversion but I was unable to sample current and voltage simultaneously (well, almost ...) even using the sample and hold properties and counting machine cycles.
This may explain that ?
Then, I went to a stereo sound board . . .
I am working on a 8051 design which does not use AC on the anode. Rather it will use a DC supply with the voltage controlled by a d/a. This way I am not time constrained (other than by dissipation of tube structures). I can step the voltage, let it settle, take a reading, proceed to the next step...

I was actually more concerned with d/a conversion time for control since I was using a serial d/a. It ended being too long so I switched to a parallel d/a and can now do four a/d conversions and one d/a cycle in 116us (c code, not assembly). This should translate to near 16 sweeps per second from 0 to 500v in 1V increments if my HV Sweep control amp is fast enough (next part to prototype).

Quote:
6P15 has no internal connection btwn K and G3 I know, I've paid
Yes, I made a jumper just for this.

What happens without the connection? Fried G3? Does this happen even with quick sweeps in a curve tracer?

Last edited by TheGimp; 20th April 2011 at 06:18 PM. Reason: add timing for 8051
  Reply With Quote
Old 20th April 2011, 06:22 PM   #18
diyAudio Member
 
smoking-amp's Avatar
 
Join Date: Dec 2001
Location: Hickory, NC
Leaving g3 disconnected will in most cases cause it to charge negative, and so lower plate current some. But g3 is not very effective in most normal tubes (other than mixer tubes), so just a slight reduction in plate current, mainly at low plate voltage. Try it and see.

Ran out of time editing my last post:
The timing issue, with only one S/H on the chip, could also be handled by using a real time clock to accurately time the sample conversions on altenating scans for plate voltage and for plate current. Assumes the line voltage is steady.

You might also try another tube, could be defective! A gassy tube would have abnormally low plate current at low plate voltage. Excessive screen grid current could be a tip-off that the tube is gassy.
__________________
Without MS, how would we live without all those hackers?

Last edited by smoking-amp; 20th April 2011 at 06:46 PM.
  Reply With Quote
Old 20th April 2011, 06:50 PM   #19
TheGimp is offline TheGimp  United States
diyAudio Member
 
TheGimp's Avatar
 
Join Date: Aug 2009
Location: Johnson City, TN
Another dissadvantage of of using AC on the plate is the non-linear dv/dt which requires a one to sample at a time varying rate in order to linearize the sweep.

One way to help is to use a voltage higher than the sweep requirement and ignore portion near peak as it changes slowly. This however worstens the voltage near zero as it makes dv/dt greater. The voltage near zero crossing is more difficult and requires fast processing to get enough data. Unfortunatly this is where multi-grid tubes have the most interesting data and requires close sampling, say 1V between samples.
  Reply With Quote
Old 20th April 2011, 07:03 PM   #20
diyAudio Member
 
smoking-amp's Avatar
 
Join Date: Dec 2001
Location: Hickory, NC
The AC sine wave sweep can be useful. It keeps the tube out of the g2 major heating region most of the time (at least for an analog tracer). But a single sweep session using digitization avoids the problem nicely.

One thing I noticed on the traces so far is that the max plate voltage is the same for each stepped trace. Usually a current limiting resistor gets put in series with the plate to limit max dissipation. Unless maybe the software is post selecting the data points. This max dissipation issue is one area where I could see using a driven plate supply and software calculating the max dissipation for each sweep.

Also very useful is an adjustable current limit on the g2 supply.

I have some Xantrex bench supplies here that have 16 bit RS-232 controls and readbacks, which I have thought of using for a programmable tracer, with Visual Basic controlling everything. Since I got the Tek 576 tracer working though, I haven't had much impetus to work on that. But for Spice tube modeling, the digitization would be essential. Another scheme is to make the curve tracer become the real time tube Spice model itself, useful for variable g2 signals and current measurements. Probably need to have the tracer software do some predictive sampling and software interpolation to stay ahead of the Spice simulation, for speed.
__________________
Without MS, how would we live without all those hackers?

Last edited by smoking-amp; 20th April 2011 at 07:26 PM.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Help interpret graphs? markusA Full Range 22 10th May 2010 04:54 PM
Plate Current VS Plate Voltage Graphs Captn Dave Tubes / Valves 2 13th February 2009 09:57 PM
The 6P43P russian tube is a EL84 equivalent ?? aldovan Tubes / Valves 3 27th May 2007 08:57 AM
FE87E Graphs philipbarrett Full Range 3 22nd November 2005 11:03 PM
Brazil, and some strange laws, not too much strange! destroyer X Everything Else 0 11th September 2004 01:38 PM


New To Site? Need Help?

All times are GMT. The time now is 04:35 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2