The AM Forum
March 29, 2024, 06:49:11 AM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Calendar Links Staff List Gallery Login Register  
Pages: 1 ... 4 [5] 6   Go Down
  Print  
Author Topic: Simple Audio Driver for 833A modulator  (Read 94346 times)
0 Members and 1 Guest are viewing this topic.
Tom WA3KLR
Contributing
Member
*
Offline Offline

Posts: 2120



« Reply #100 on: November 07, 2005, 05:26:05 PM »

HI Bobby,

No you're not missing anything.  You are correct.  But the guys have a pair of 833's to modulate with.  Staying in AB1, theywill still have enovhg audio.  We coulddo what you say.  My circuit does hit the rail on the negative swing. but the transient response is good.  By the time the waveform comes up out of the cut-off region, everything is fine.
Logged

73 de Tom WA3KLR  AMI # 77   Amplitude Modulation - a force Now and for the Future!
nu2b
Guest
« Reply #101 on: November 07, 2005, 08:27:36 PM »

Tom, I know, I was just trying to get the most out.
Yep, my model shows the loop slamming the rail but the TL082 seems
to recover OK. I also incorporated Franks suggested phase splitter
using a 2222A. This allows true line level +/- 1.1 vpk input.
Just split the collecter and emitter loads into 2 resistors each
to get a 6 db drop, then the input handles line level plus 3db peaks.

Also, strap the grids to whatever supply (+15,+24,+75...) you have
to get a little more plate and grid monkey swing and also to move
the bias node up from 0.7v (.85,.94,1.45...) More monkeyswing.
This means that you can handle a little more audio in.

Do you know if TL074, TL084 are true rail-to-rail?
If so, then you might be able to get rid of the -5V supply.
My model starts to crap out when the TL084 gets close to
the rail using 0V rather than -5V supply.

Send me your E-mail and I'll send you the mod on how to turn
a 12AX7 into an 833A. I'll try to post it here when I have more
time. Oh, did you ever get a chance to test drive the Class-E
program I sent?

Regards,
BobbyT
Logged
Tom WA3KLR
Contributing
Member
*
Offline Offline

Posts: 2120



« Reply #102 on: November 07, 2005, 09:32:57 PM »

Bobby,

I’ve made 3 pdfs of simulation results of the situation that you are concerned about. 

Sim1 – Shows both of the cathode driver output waveforms.  The simulation has a load that roughly models the current draw and cathode voltage relationship of the 833s.

Sim2 – I’ve overlayed a matching sine wave to the driver output waveform.  The driver output waveform limits at 109.2 Volts dc.  When the sine wave limits at 109 Volts, the loop will go out of regulation.  At the peak of the input waveform, the driver op amp’s differential inputs separate by 265 mV.  I have circled the region where the loop starts to recover and come back into regulation.

Sim3 – This is a blow-up of the region circled in Sim2.  Use the point where the “test” waveform crosses the output waveform as a reference for recovery time start.  This is 405 uS on the graph’s X axis.  The cathode output waveform snaps back downward with a very well behaved response with only one cycle of ringing.  At 420 uS on the graph, the output waveform has reconverged with the “test” sine wave.  This was 15 uS, to recover.  At this point the voltage is 104 volts dc.  Still in the cut-off region.  Another 15 uS. to go to get to 100 V. (-100V.).

The pull-up power supply voltage could be raised so that the recovery is at a more negative grid-to cathode voltage.  But I don’t see that as being necessary as of yet.  The loop has been compensated for a wide range of FETs.  But it would be best to use one specific type of FET with the lowest capacitance necessary and re-do the compensation study for the best performance with that FET.  But I think that you will agree that the performance is not an impossible situation with terrible outcome.

So I feel that it is not necessary to make the waveform swing any farther “negative” than it already does.  The simulations show that it is not necessary to mimic the whole voltage swing of the old grid driver transformer.

* FETsim1.pdf (22.14 KB - downloaded 517 times.)
* FETsim2.pdf (28.39 KB - downloaded 534 times.)
* FETsim3.pdf (12.33 KB - downloaded 555 times.)
Logged

73 de Tom WA3KLR  AMI # 77   Amplitude Modulation - a force Now and for the Future!
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #103 on: November 08, 2005, 08:55:44 AM »

hi Guys,
Tube curves show plenty of plate current at 3 KV and grids at zero. This is where Tom Vu and I started. 70 volts on the cathode should cut it off.
Tom klr I still like eliminatinting the first op amp and eliminating the extra delay in the negative phase. Yup a fet or a 2n2222 would do the job. KISS. The only other way around it is use the spare amp as a noninverting the match the phase delay.
Pulling the grids positive would be a last resort to get more current. Save that thought.
074 and 084 are not rail to rail but the fets need about 4 volts to turn on so -5 negative rail should be fine to yank them off.. -12 good also.

Tom KLR extra resistors in input only balance the input Z eliminating possible offset voltages. yes your circuit will play but I'm picking fly crap out of the pepper shaker for best possible performance.
Feedback in the source resistor may have start up transient so you have to rely on your soft start at bias pots. This means the driver needs to be switched with the
rig to avoid mod transformer jumping up going into TX. cool stuff
Logged
nu2b
Guest
« Reply #104 on: November 08, 2005, 12:10:25 PM »

Tom, as Frank said, Cool Stuff.

Good on the graph details. The transients look good.
I can't wait till somebody puts one together.

Here's the 12AX7 starting model I promised.

*Vacuum Tube Triode (Audio freq.) pkg:VT-9 (A:1,2,3)(B:6,7,8)
.SUBCKT X12AX7 1 3 4
B1 2 4 I=((URAMP((V(2,4)/85)+V(3,4)))^1.5)/580
C1 3 4 1.6E-12
C2 3 1 1.7E-12
C3 1 4 0.46E-12
R1 3 5 50E+3
D1 1 2 DX
D2 4 2 DX2
D3 5 4 DX
.MODEL DX D(IS=1.0E-12 RS=1.0)
.MODEL DX2 D(IS=1.0E-9 RS=1.0)
.ENDS X12AX7

Also,
Here's a first approximation for an 833A Ip vs Vgk and Vp curve.
with cutoff at approx -70V and drawing about 0.9 amp at 3KV.

  Ip= (Vpk/42.9) + Vgk)^1.5)/650

Folks could change the model to reflect this, add the proper tube capacity and
maybe change R1 to about 400 ohms to approximate the grid requirements.
Is this fun or what!

Regards,
BobbyT
Logged
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #105 on: November 08, 2005, 12:25:07 PM »

Imagine a 3 by 4 inch board with 2 fets hanging off it with a little transformer supplying power driving a pair of 833s. We are debating resistors. Imagine all the hardware that can go to the scrap heap.
Logged
nu2b
Guest
« Reply #106 on: November 08, 2005, 12:33:05 PM »

The next thing to do is come up with an old buzzard design to go along
with the old buzzard tube. That is, just use solid state triodes like transistors
and FETs along with zeners.

"We don't need no stinkin' opamps and voltage regulator chips!"
Logged
nu2b
Guest
« Reply #107 on: November 08, 2005, 12:41:34 PM »

Hey Tom JJ,
You can be the first to build this!

Just build 1 channel and convert your CakePan rig to series Cathode Modulation.
Logged
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #108 on: November 08, 2005, 03:13:13 PM »

Hey Tom JJ,
You can be the first to build this!

Just build 1 channel and convert your CakePan rig to series Cathode Modulation.

Oh, but it's already being done here in the shack here Bobby. My 6AQ5 series modulated by another 6AQ5 sees occassional use on 75M... Grin  I like cathode modulation.  2W output and clean 200% easily.

http://www.amwindow.org/tech/htm/series.htm

So, what's the next step here? Maybe Frank, you and BobbyT can come up with the simplified version based on discussion today. I'm glad I'm not ready to build it now cuz I wouldn't know which to pick.

T

Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
Tom WA3KLR
Contributing
Member
*
Offline Offline

Posts: 2120



« Reply #109 on: November 08, 2005, 03:44:53 PM »

Frank, you are right that a current feedback version would have start-up problems.  I hadn’t thought about it yet.  The loop will be asking for zero bias to get current.  There probably will be grid current in standby, if there is a plus supply on the grids, equal to the desired 50 mA cathode current the board would be set up for.  I don’t see any grid current or grid dissipation specs in my book but there has to be.  So the current feedback version needs to be a little more complex than the present “voltage output” version.

The voltage output version and 833’s are happy all the time.  The driver board doesn’t need to know what mode the modulator is in.

Not previously stated, the voltage driver board can be used with full AB2 and the positive grid bias with no changes to the board design; that is not precluded.  It’s just that you need to adjust the bias pots, have a higher pull-up voltage, and have larger heat sinks for the increased FET dissipation.  Since the driver has fixed gain, more audio drive would be needed for greater output voltage swing.  For the design as it is, with the cheapo supply I published, you could install a second set of diodes and a big filter cap feeding an additional 7818 for the grid supply.  This would help somebody like Mike W3SLK to get his 810’s from AB1 330 Watts I stated before, to about 415 Watts, over the hump I think, for him.

About the 210 nS delay in channel 2 – It hit me that to combine both of the outputs together for analysis, all I needed to do was grab an ideal op amp and I set it to a gain of 10, duh.  I had to run both channels at a level below clipping; both channels are putting out sine waves.  I ran the transient simulation at 10 kHz. and measured one channel’s distortion and the distortion from the combining op amp’s output.  I measured 0.204 % on the cathode2 output.  Coming out of the combiner op amp I measured 0.144 %.  You have to remember that the push-pull “combining” attenuates the even harmonics.  I think that we are fine with the technically offset timing.  This is just a voice product.

I measure the distortion by asking for the FFT of a waveform.  I then measure how many dB down the second, third, fourth and fifth harmonics are.  I plug those numbers into a THD calculator spreadsheet I made.  It has been a very good tool to use.

So then I put in the 2 resistors ahead of the op amp inputs (100k in parallel with 1k = 990 Ohms in this case) you are talking about Frank, and re-ran the simulation.  I got 0.386 % on the cathode2 output and 0.380 % on the combiner output; a surprising result.  It is worse!

I never put those in anything I did except a thermocouple amplifier, which had much, much more input bias current that these bi-FET input op amps.  The TL072 input bias current at room temperature is 65 pA.  With a 1k series input resistor this is 65 nV.  With a gain of 101, this is 6.6 uV on the output, totally insignificant in this application.  You also need to remember that the input offset voltage spec. on this part is 3 mV.

I also thought about the fact that we could put the unused op amp in channel 1 as a voltage follower.   This is a possibility, but again more complexity for little gained.  We may need it for the current feedback version.

So I think that the circuit still stands as is.

Logged

73 de Tom WA3KLR  AMI # 77   Amplitude Modulation - a force Now and for the Future!
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #110 on: November 08, 2005, 03:47:39 PM »

Tom's circuit will work, go for it. My suggestion of a phase splitter would be not make much difference. I was just being anal. But you have to be anal to get the last couple tenths of a percent.  
Be the first on your frequency to cathode drive your modulator.
Linear that is, I've been driving driving the cathodes of my pdm rig since '83 and never blew a single fet.


Logged
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #111 on: November 08, 2005, 03:57:09 PM »

Tom KLR,
Yup I agree. Interesting on the input resistors. Yea input current is so low the offset will be in the mud. Not like the old days of bipolar amps.
I think we have a very good base line, and just need a good heat sink for growth.
It would be cool to keep the grids at ground. Maybe a negative reference below the FETs since audio is ac coupled in. That would be cool to provide a path back to the high voltage supply....hmmm maybe a little positive grid bias is safer.

SOOOOOO who is first I'm going back to my new RA6830 project.
Logged
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #112 on: November 08, 2005, 04:08:53 PM »

Tom KLR,
One last thing. Rig in RX with high voltage on the modulators. Will there be enough tube leakage to soar the drains of the fets. GG linears always put 50 Kohms across cathode relay contacts. Do we need a high resistor across each FET or will the pull up hold it low enough. Also don't want to back feed the pullup supply...maybe a 1n4007 series diode ? I'm not sure what happens when the cathode is an open circuit.  I would suspect that voltage may get high since there is no leakage path.
The circuit may be off so no current flows in RX. Just a safety thought a SMS guy.
Logged
nu2b
Guest
« Reply #113 on: November 09, 2005, 09:16:18 AM »

Tom, Frank
What if you rebiased the opamp reference nodes higher to eliminate
the negative supply.

For instance:
Use +15V
Ground the old -5V opamp supply terminal.
Bias the (-opinput) to +7.5V with 2 100k res, R1,R2
Insert R3- approx 11k to ground- in series with the existing 1k to (+opinput)
Both opamp inputs are now at +7.5v nominal.
Bypass R3 with about 100uf to ground to restore original ac loop gain
and establish low freq corner.

This should give about 70 volts at the cathode as before.
Make part of R3 the bias adjust pot.

Can use higher than +15V but if higher than +18V use zener to clamp the gate.
This is not soft-start, so AC supply for +15V should be powered from the same
source as the 833A fils. Might also provide a 3KV plate lock out
in the absence of +15 and +150 pullup supply.

If +150V supply has a bit more extra mils, the whole thing can be run from one supply

Also, what if a separate dc bias loop were added to the grid terminal.
This would provide a tracking input to follow the grid supply.
This way if the grids were biased to a higher voltage, the idle current
could be made to track. Just feed back 51mv per volt of grid bias to the
bottom of the 100k (-opamp) using a 10k and 510 ohm res?

What's your opinion on this working?

Regards,BobbyT
Logged
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #114 on: November 09, 2005, 11:51:37 AM »

I would think you could raise everything higher as long as the op amp negative rail voltages goes low enough to turn the fets off. I'm not sure it will be as quiet in  the hum department. The op amp negative rail at ground never seems to work as well as a negative supply.  I would leave the grid out of the loop for now as it complicates things. I prefer to see the grid at ground for starters.
Logged
nu2b
Guest
« Reply #115 on: November 09, 2005, 12:48:34 PM »

That's true Frank,
You loose the inherent opamp CM power supply rejection.
However, the ps bias splitter is still inside the closed loop,
so (for the +15 V single supply situation), you could go to
+15.5V and have the output only change about 2V.
This is common mode for both cathodes such that
the push-pull output is not disturbed.

If you use a 30V PS the effect is halved
(also get full opamp output sink-source performance to drive FET cap at high freqs)
 You could also maybe pump this in as a reference to the other half and cancel the effect.
Maybe not even use a regulator, just a Zener for the 15V bias splitter.
Maybe all of the above?

Regards, BobbyT
Logged
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #116 on: November 09, 2005, 03:01:39 PM »

I would prefer to keep it simple. Fixed bias on the grids would be compensated by the cathode voltage. Floating op amps with  offset bias is just drift looking for temperature change. 
Logged
nu2b
Guest
« Reply #117 on: November 09, 2005, 04:30:03 PM »

"Duh-brainfart"
The bias is definitely not in the loop.
That's what I get for trying to type and think at the
same time.

The other comments were hopefully on target and were an attempt to KISS.
Can always try it both ways.

BobbyT
Logged
nu2b
Guest
« Reply #118 on: November 12, 2005, 11:53:34 AM »

Folks,
Here's a better approximation for an 833A Ip vs Vgk and Vp curve.
with 50ma bias at -70V and drawing about 0.896amp at 3KV, cutoff is about -82V.

  Ip= (Vpk/36.6) + Vgk)^1.5)/828

Another thought.
I'm not sure what happens with an open circuit cathode but
maybe Frank's 50kohm pulldown should be used directly at the833 cathode.
This means if the fuse blows, the tube will bias up to about 82v.
This way the grid voltage breakdown would not be exceeded.
If breakdown at 500v was reached does that mean that Vg=Vk=zero bias  and maybe
0.7amps of grid current is drawn? Does anyone know?

Regards,
BobbyT
Logged
Tom WA3KLR
Contributing
Member
*
Offline Offline

Posts: 2120



« Reply #119 on: November 13, 2005, 08:28:13 PM »

Op-amp re-biasing to eliminate the – 5 Volt supply:
This could probably be done with another op amp choice, but I am not going to pursue it.  Rail-to-rail input and output op amps were designed for the battery-run equipment.  I think that we would want one that handles more than 5 Volts.  They probably exist but I am not going to do the research.

What we would need is common mode input to ground and the output would not need to go to either rail.

Yes you would have to bias the feedback divider at R17 to some positive voltage.  I feel quite uncomfortable using a high impedance source at the bottom of R17 though.  The AC and DC gain will be different.  If I were to modify the design, I would use a 78L05 regulator at the bottom of R17.  Then we are eliminating the 79L05 and putting in a 78L05.  So I don’t see an advantage here.  2 diodes and a 7XL0X i.c. are cheap.

I was looking at the 833A curves again today and I plotted the class B load line from the tables and the current output is quite linear until + 50 Volts bias.  If you add a positive voltage regulated grid supply, I still don’t see any  need to do any other bias regulation than setting the bias voltage in the present circuit.  The grid to cathode bias would be defined.

I did an experiment on the bench the other day with a 6J5, grounded grid.  Plate supply was + 360 Volts d.c.  I dug out my high voltage probe that goes with my RCA Senior VoltOhmyst.  The probe is 1100 MegOhms.  I saw 138 Volts on the cathode of the tube.  Scaling up 2 orders of magnitude for the 833, I can’t imagine that even with 1 MegOhm on the cathode that the voltage would climb near the breakdown rating of the FETs.  But this is just a guess.  We will have 10 k to +125 Volts anyway.  So I think that we are o.k. in the situation of the triode as a possible source of high voltage.

The fuse was an idea in case the FET shorted.  The 833 can conduct over 2 Amps.  At first the current would follow the load line.  But when the tube and transformer sit there steady, wouldn’t the tube operating point be vertically straight up the 3000 V line to where it intersects the extrapolated grid bias line?  The fuse rating should cut down some if incorporated.  Looks like the tube would sit at about 0.85 Amps considering the 1 Ohm resistor in the FET source.  Of course, if and when the fuse pops, that might make quite a voltage transient in the mod trranformer; a bad thing.

I don’t know what the grid breakdown rating is so if it is 500 Volts, perhaps a resistor past the fuse is good.   If the cathode was floating, there is no current path for grid current.

Perhaps there should be no fuse except in the mod. plate supply.


A current-feedback version cathode driver:

I have taken a few stints at the current-feedback version driver.  The situation is not easy.  I have a working circuit with about 1 % distortion.  But I am having problems with the linearity at low currents.  The source resistor that was 1 Ohm in the voltage feedback driver had to be cut down to 0.1 Ohm so that it didn’t interfere with the gate voltage.  Now we are dealing with 5 milliVolts for bias current reference.


Bobby,

I ran your CE04 program this weekend.  I like the table showing the component currents and voltages.  I had thought about doing this in my Class E spreadsheet but didn’t want to take the time to develop rules of thumb checked at various impedances and QLs.  I haven’t compared your voltage and current values to my old PA simulations yet.

I like the coil section where the actual Q of the described coil dimensions is estimated.  I have never seen this before.  Have you ever measured any real coils against the calculated Q?
Logged

73 de Tom WA3KLR  AMI # 77   Amplitude Modulation - a force Now and for the Future!
nu2b
Guest
« Reply #120 on: November 14, 2005, 01:03:43 PM »

OK Tom,
I was just trying to nail any last minute "gotchas" or domino failure modes.
Was concerned that maybe old gassy 833 pulls with maybe 5 mils of leakage
would not cut off like a hard vacuum tube would. This might pullup to the point either of Vgk breakdown or filament transformer breakdown.
Also was thinking of tube "burps" that APE mentioned.

On the ClassE program:
I don't have a Q-meter so haven't made measurements. The Q estimate should be valid. Wheeler's estimate is too optimistic, Butterworths is closer and seems to match the ITT handbook estimate. Try the help button to get you to the references to the literature.
I put in the component voltage and current so folks could better estimate tuning and loading cap requirements. Maybe just a breadslicer would do rather than a vacuum var, etc.

BobbyT
Logged
nu2b
Guest
« Reply #121 on: November 15, 2005, 12:46:11 PM »

Tom,
Your power supply with one transformer that has a dual primary is
a great idea to save space and keep KISS going.

I have a question on how to figure the rating though.
Normally the transformer should be derated to 70% for cap input usage.
If we start with a 10 VA transformer, we're down to 7VA.
Since the transformer was originally wound so that both primaries supported the rating,
does this mean that with only one primary excited, the single primary can only support
3.5 VA without overheating?

Maybe we could model this. See below info from a spreadsheet generated a while back when
I was looking at how to model this kind of stuff.
The values are for full load 40 deg rise. So for room temp divide resistance by 1.2.

BobbyT
---------------------------------------------------------------------------
Simple Iron-Core Transformer Model with winding loss but excluding core-loss.   

Select         Calculated
            
PRI Vrms   120   PRI-VA=   9.53
Freq-Hz        60   PRI-Irms= 0.079
      SEC-VA=   8.10
SEC Vrms     36   Loss watts=1.43
SEC Irms   0.225   Rpri-Reflected-Tot-ohms= 227
      Eff Vin-rms= 102.0
            
%Efficiency  85           Nsec/Npri=     0.353
      Sec-Vrms OpenCkt  42.35

      Rpri=   113.3
      Rsec=   14.19
Lpri-HY           8   Lsec-HY=     .997
            
      CT Rsec=    7.06
      CT Lsec-hy=  0.249
            
      Pri-Reactance   3015.93
      Exciting Current   0.04
Logged
Tom WA3KLR
Contributing
Member
*
Offline Offline

Posts: 2120



« Reply #122 on: November 15, 2005, 03:33:46 PM »

Bobby,

I went back and looked at my power supply simulation.  I have a 20 mA. load for the 78L15 supply,  15 mA. for the 79L05 supply, and 15 mA. for the +127V supply.  The currents are a little over-estimated.  I came up with 864 mW true power for the low voltage secondaries and 1.435 W for the 120 V winding.  This is a total of 2.3 Watts true power. 

The SE Asians are pretty good at pinching each parameter in their transformer designs for lowest cost, but I really don’t know what the transformer’s limiting factor is.  My guess is that each primary winding is good for more than 3VA/0.85 = 3.53 VA and that the core is the limiting factor.  But assuming that each primary is only good for handling 3.53 VA, I may be a hair over the spec. I think that a 6 VA transformer would be fine in reality.  Remember the loads are a little over-estimated and the service is PTT.

But if you feel uncomfortable, go to a 10 VA unit.  Here’s a Mouser 10 VA unit – a good old Fred Hammond p.c.b. mount however, 546-183G36, 10 VA, for $12.15 .

Added later (sometimes I get too concise in an attempt to save time) -

So total secondary power is 2.3 Watts.  Using your 85 % efficiency, 2.3/0.85 = 2.706 Watts at primary.  ASSuming 0.75 Power Factor at the primary ( a very reasonable number) this gives 2.706 W/0.75 PF = 3.6 VA.  The transformer was sized orginally with the idea that only one primary winding would be used and that the total load was a little under half the rating.

Logged

73 de Tom WA3KLR  AMI # 77   Amplitude Modulation - a force Now and for the Future!
nu2b
Guest
« Reply #123 on: November 16, 2005, 12:35:07 PM »

Tom,
It wasn't so much concern as oldfart conservatism. It seems that when you get
started on an interesting design problem like this, the juices get flowing and
new thoughts and ideas keep popping up such that the problem gets engineered
to death. Always seems that new stuff wants to get added in and the original
design iterates toward a universal solution rather than a point design. I'd keep the power supply
separate from the board design for this reason.

Closing thoughts before this gets beat into oblivion:
 
1) Change from individual bias pots to a master bias pot plus balance pot.

2) Add series res from each 1 ohm source resistor to a 3 position switch to
provide cathode A, cathode B and total modulator current metering.

3) And finally, howsabout adding an old buzzard 6E5 magic eye to the center of the front
panel below the 833 viewing cutout for "just because" reasons only.

Regards,
BobbyT

PS: The transformer spreadsheet didn't print too well, so I'll send the excel
file to you when get a chance.

Also, I've been trying to model assymetric (pos peaker) modulation and came up
with a KISS symmetrical line level compressor which if offset slightly should
give the effect. Maybe could be used as neg peak control also. Needs more
thought. Maybe you have some modeling thoughts on this.

Also again, were you able to create a tube macro/subcircuit in SwitcherCad?
How about a modified depletion mode FET with gorilla characteristics?
Logged
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #124 on: November 16, 2005, 12:39:46 PM »

Hey guys ,
The spare op amp would be a great negative peak limiter with the addition of a couple resistors a diode and maybe a level pot. Compression inside the chassis sounds dangerous. fc
Logged
Pages: 1 ... 4 [5] 6   Go Up
  Print  
 
Jump to:  

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands
 AMfone © 2001-2015
Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Page created in 0.064 seconds with 19 queries.