Time to work on stuff - huge audio improvements.

[Opcom]

Last week, I spent a lot of time working on my station. How amazing! what did it take?  My former employer decided to terminate the applications lab. Unfortunately that was my lab.

I am 'employed' until DEC 31, but don't have to show up to work and will also get my vacation time pay, severance, and the bonus I share in.

They were pretty nice about all of it actually. Although it is regrettable, there are no ill feelings. I have enough visibility to see the business decision. How odd though. I have never been let go from a job before. Ever.


This has left me with time. I can't stand to lie about all day, or to do nothing; and I have already seen the internet.. So to the point of this post:



For the last month or so, I have been off the air, the station;s audio section has been tore down and I have had no time. It is ready now, to go back on the air. Because of all the information that was figured out, and how the audio drive is working perfectly now audio-wise, it's worth a detailed report of what was done.

It was all torn down from the mike to the modulator grids. Many tests and measurements have been made. The issue with not getting 100+ percent modulation was traced to the speech amp, and the mismatch between the 3-500Z grids and the speech amp plates has been discussed.

1.) Not enough voltage to the input of the modulation driver transformer

2.) really bad regulation of the drive voltage to the modulator grid over the half-cycle. It may be seen on a scope and corresponds to modulator grid current variations as seen on the scope using a shunt and a balanced probe arrangement in series with one of the grids. The Z varies a lot and this was distorting the waveform. The modulator was loaded with resistors for this so it would not be done with RF everywhere It was an interesting lash-up.

3.) Not enough power from the speech amp to drive the modulator to clipping. The speech amp clipped first.

4.) Good sounding but horrible mike arrangement and no simple PTT. Going RX->TX was more like starting up an aircraft each time. That gets old

5.) Horrible mike preamp with wrong frequency response (RIAA curve) not enough gain. One had to yell at the mike.

6.) No compressor or clipper, so a low average modulation to avoid splatter, assuming the occasional peak

7.) Wrong speech polarity, or is it? No instant way to swap for A/B test.







What was done for issues 1 through 3 is to use a 150 Watt tube type public address amp. A Stromberg Carlson AP-80 was used. This amp has four 6550's in pp-par. and an output winding with 70V and 25V center tapped.

The problems were clear enough. An irritating issue is that there is no audio transformer commonly available for driving a pair of 3-500Z's in AB2/B push pull. The usual transformer is for 250T's, 450T's, etc. The 3-500Z takes about half the voltage and twice the current compared to those other tubes. 1/4 the impedance. The common modulator driver transformers, like the CG-512 and its smaller cousins, usually have a set of input taps for 50-500 Ohms. The 'direction of tapping' for 3-500Z's is the opposite; a lower turns ratio not a higher one.

Rather than spend more time trying to determine the best match and polarity, a switch panel was made and gives the following 'RMS' outputs. This is just a selection of 4 taps. one side of the 70v output is common, and one of the other 4 taps are chosen. Note the other possible impedance relating to the 25VCT winding, 4 Ohms, is not available, but it does not matter as there is a 22.5V tap.

70V - 32.7 Ohms

47.5V - 15 Ohms

35V - 8 Ohms

22.5V - 3.375 Ohms

The impedances seem low, and they are quite low, and the amplifier is at least 3-4X the power level needed. An output of 70V is the max, but this is >2x the voltage needed to make the modulator clip when driven through the existing CG-512. The 'right' tap is the one that gives the lowest voltage but can still drive the modulator to clipping.

Because of concern that the amplifier could operate with too small a load, a 10 Ohm resistor was put always connected across the 25VCT winding taps. Via some spreadsheet math, this makes the resulting loads acceptable for a proper match to the PA amplifier:

70V - 56 Ohms

47.5V - 25.8 Ohms

35V - 14 Ohms

22.5V - 5.8 Ohms

The 10 Ohm resistor eats 62.5W when the amp makes 150W. This leaves 87W available for modulator drive, and it improves regulation a lot by reducing the impedance variation seen by the PA amplifier over the half-cycle.

In previous testing, it was seen that the previous seech amp, a Stromberg-Carlson 30W hi-fi amp, worked best when the 32 Ohm tap was used. There was no higher tap, but a 1:2 step up transformer placed after the amp cause a lot of overloading of the 30W amp, it could not supply the current at the higher Z and ran short of power. It was almost enough, but it was not desired to operate things at the edge. (Or to use a non-rackmount hi-fi amp with costly 7027's for this duty)

The CG-512 is rated for 40W and is one of the largest driver transformers sold from catalogs. It was luck that it's what COL Tucker chose. The peak current going into it is twice the amount that there would be for the higher-Z tubes it was made for, but its large size seems to have offset any problems so far (like burning out).

It would be FB to have a custm modulator driver transformer wound on the laminations from an old 40-50W audio outut transformer but that situation is not readily available except by special request to one of the transformer rewinding gurus we are lucky to know through this BBS.

So, that illustrates the value of having an amp making 3-4x the necesary power, and steps taken (swamping) to improve regulation despite the PA amp being push pull AB1 beam tubes (triodes are usually better for drivers of class B stages).

This was the most time-consuming part of fixing up the shack.





Next, issue 4, 5, 6.
The Altec mike mixer and Beringer compresor+clipper have finally been put to use, and a rummage through the microphone box showed that every "professional" mike with an XLR connector was bad. An $18 one from China was found. It works OK, and is low impedance. The Altec has a bunch of high impedance to 500 Ohm transformers on it. These are Octal based. The one for the channel the new mike is pluged into had to be removed and in its place, an Octal tube socket base with jumpers was inserted. The mike goes into the Altec mixer which acts as a preamp. it has a meter on it. This then drived the Beringer compressor, MDX1600, I think it is. There are a lot of fiddly controls on that thing. one thing is that it has a clipper but the clipper seems to miss the first attack. This was disappointing, because what else is a clipper for but to limit the audio and clean up after the compressor? Anyway, over 23dB of variation in the mike signal input results in a constant out out, so it is acceptable. I still need a boom or holder for the mike. It's a hand mike like a singer would use. Not sure where to find the clamp and preferablya flexible gooseneck that could be boltd to the rack ot something.

The 'controls' issue is not done. needs more figuring out.






Issue 7 
- was done by using a DPDT switch between the PA amplifier's selected output and the Tucker's modulator input. It's in the same panel as the impedance selector. Above that is an old RCA broadcast panel with a 3" meter set to indicate the voltage from the 70V winding of the PA amp's output. This way it can be seen always if that amp is ever being driven close to clipping. It's a 0-3VDC meter. A bridge rectifier and voltage divider was used. No matter what, it always sits on the 70V output of the PA amp. A 100uF cap helps keep the pointer up since it is peaks that are important.






The last thing -the controls from #4.

There is no way around this. It is ghetto to have to throw 3 switches and turn two knobs going from RX to TX and so forth. The answer is relays. Lots of 12VDC relays, which are here, in a box. No need to describe what they will control, it is the typical keying up, muting, and the rest. I have a scheme to allow semi-break-in on CW for my friend Dennis when he comes over. The challenge is the antenna relay, but that should be do-able with one large one with good insulation, nothing exotic, 600V/20A maybe shoudl do OK for the RF and last a long time. A PITA to wire up the connectors and all, but that's how it goes.

It'll have to be more than a little box, probably a panel. The reaon is there are 4 receivers, one transceiver to be used alone or as a receiver, and one transcever to be used alone, with the linear amp, or as a VFO/driver for the Tucker TX.

Now, I am using those 4-position 'MFJ" RF switches. they are OK, but there are too many and it is very messy looking. A patch bay would be allright, but that can be messy too, and fiddly always screwing and unscrewing, and a mistake could be disastrous. Like transmitting into an R390.





OK well I have the rest of the month off, and there is little point in looking for a job until after JAN 1 2014, unless I might want a temp at a department store, but all the young folks have taken those.

Below:
I have been the cat for a long time. As of last Monday, It's my turn to be the arm.

[John K5PRO]

Sorry to hear they dropped you like a rock, Patrick, but you sound like you have a reasonable attitude about it all.
I sometimes wish I had a lot of spare time again, instead of constant work, to fix up so many radios. I repaired a troubled TR7 Drake a few weeks ago at least. I was in Dallas last week at Continental Electronics, then got out of there before icemageddon hit Friday. I am returning possibly in a few days.

The audio improvements are the right way to spend your time now!

[Opcom]

Thanks John,

Next time you are in Dallas, gimme a call. I owe you lunch or dinner for that cool frequency monitor and filter choke you gave me.

Oh yeah we had the ice horror. The only people driving were emergency vehicles, a very few 18 wheelers, and the foolhardy public.

My backup diesel set has a coolant heater that is 1KW. I have always considered it to be too big/wasteful and run it from a 1KVA control transformer set up for 2:1 voltage step down, but the last couple nights I switched it over to the full 120V. Also searched the whole lab to find an old 250W magnetically attached crank case heater and stuck it on the bottom of the fuel tank to keep it from gelling. Even with this the fuel temp was just 33 degrees when it was so cold.

We don't see this 15 degree weather much and most outdoor equipment isn't really ready for it. There have been several large power outages. I don't know what they do up north when they have this kind of mess every year, or how that infrastructure holds up.

[N2DTS]

I found it usefull many years ago to build a station control rackmount box.
I had some nice small rack mount boxes from work that had lots of holes already for coax connectors and terminal strips on the back. I made up a new front panel, and put switches on the front for 7 receivers, 4 transmitters, a switch that selects which rig triggers the T/R relay, and a mod monitor takeoff (small wide spaced variable cap).

The T/R relay is a big ceramic base contactor mounted on rubber isolators, and I have three 3 pole double throw relays in there, with many normaly open and normaly closed contacts out the back.

I also have some coax connectors to bring out an antenna tuner/watt meter, can be in the TX only path, or on RX and TX.

It has inputs for 4 antenna's, one goes to the dummy load.

That makes for a LOT of coax connectors coming out of the rack, plus all the PTT and mute connectors, but makes hooking something up a snap.

I have another box with all the holes in it, and might have the right size rack panel to make a nice front, I paint them black and label with a P touch, works well.

If you are interested, I can give you the spare box, plus I think I have a load of terminal strips to fit on the back, and have some rubber standoff isolators, and even a bunch of 3 pole 120vac relays and sockets I think.

The box (about 6 in X 12 in?) has been sitting around on the shelf long enough...
The top comes off for easy access....

Let me know if you are interested.

[steve_qix]

Hi Patrick,

I'm really sorry to hear your job went away.  Hopefully there are other good and interesting opportunities in your area!!!!!!!

Just a quick note on the audio driver - eliminate the transformer one way or the other.

If you want to use tubes, use a cathode follower direct coupled driver.  Solid state is better (a source follower), but the tube designs work if you put feedback around each leg of the driver itself to reduce the source impedance.  Either way, you need a low (very low, actually) impedance source for class B or AB grids, and as a significant added bonus - since there is no phase shift introduced by the driver transformer, you can put a whole lot of negative feedback around the modulator itself including the modulation transformer

Conditions are good on 75 meters - maybe will here you on and say hello.

[Opcom]

Let me know if you are interested.

PM Sent! Thanks!

[Opcom]

Hi Patrick,

I'm really sorry to hear your job went away.  Hopefully there are other good and interesting opportunities in your area!!!!!!!

Just a quick note on the audio driver - eliminate the transformer one way or the other.

If you want to use tubes, use a cathode follower direct coupled driver.  Solid state is better (a source follower), but the tube designs work if you put feedback around each leg of the driver itself to reduce the source impedance.  Either way, you need a low (very low, actually) impedance source for class B or AB grids, and as a significant added bonus - since there is no phase shift introduced by the driver transformer, you can put a whole lot of negative feedback around the modulator itself including the modulation transformer

Conditions are good on 75 meters - maybe will here you on and say hello.

I thought about  pair or quad of 6AS7's or 6080's. It would not be too hard. There is room inside for a panel with some tubes like that. I like these two articles on class B grid driving.

TX is operating but There are a few things yet to do before I can go on the air, but I'd like to try 75M. It has been doing well. I think I'll just lash up something temporarily as a control until I can build a nice controller.

[DMOD]

Last week, I spent a lot of time working on my station. How amazing! what did it take?  My former employer decided to terminate the applications lab. Unfortunately that was my lab.

Man, that sucks. I could never understand why they do such dastardly deeds around Christmas time.

Pat, is this the upgraded or former Turner rig you're working on?

Phil - AC0OB
 

[N2DTS]

Yes, tis the season to be laid off.
That is when many companies do it, after you run the charge card up for gifts and so on...

The hot setup is to drive class B grids directly with some solid state stuff, or use AB1.
I think I have a bunch of 6080's if you want to play with them.
I pulled a bunch out of some military stuff that was in the trash.
I use one in my 4x150a mod deck in the screen voltage regulator, and its been the same tube for the last 30 years...

I will include some in the package.

[steve_qix]

The general concept of the cathode follower driver is a good one.  You can make a *vast* improvement in the circuit by not using the transformer the old circuit shows.

Just make a resistance coupled amplifier circuit instead and drive the 6080 (or whatever) grids with that.  Also, put a loop of negative feedback around EACH driver back to its respective resistance coupled amplifier.  You will need 1 inverted amplifier stage somewhere in each feedback loop, else the feedback will be positive 

[Opcom]

Pat, is this the upgraded or former Turner rig you're working on?

Phil - AC0OB
 

This is the "COL Tucker" transmitter, the same monstrous beast.
Upgraded is becoming more of a name for it as these annoyances are worked out.

The hot setup is to drive class B grids directly with some solid state stuff, or use AB1.

The rig is socketed for 3-500's. The only other things to put in would be 4-250's or 4-400's Those are costly and I don't have them, and I would have to add a screen supply, and revert the modulator bias supply back to high voltage, which is not practical right now.

I think I have a bunch of 6080's if you want to play with them.

I will include some in the package.

Thanks, I have 1 or 2 well-used ones and don't know their condition. Greatly appreciated!

The general concept of the cathode follower driver is a good one.  You can make a *vast* improvement in the circuit by not using the transformer the old circuit shows.

Just make a resistance coupled amplifier circuit instead and drive the 6080 (or whatever) grids with that.  Also, put a loop of negative feedback around EACH driver back to its respective resistance coupled amplifier.  You will need 1 inverted amplifier stage somewhere in each feedback loop, else the feedback will be positive  

I am sure that is 100% right.

I spent the day thinking about it and adapting the 138-139 file "class b driver". A preliminary idea is attached but not complete.

If I treat a 6080 driver like the audio section of the RCA BTA-250, then adding feedback from the modulator plates back to the circit of the cathode follower ought to be OK.

There is no voltage gain in this diagram, but one advantage is having no sensitive audio stages inside that transmitter which has been buggy.

Some voltage step-up from the 'line' can come from the existing driver transformer, and only a small amp would be needed to drive it. It's also making a convenient way to set up the feedback the same as with the RCA BC transmitter.

As it is set up, it makes a 1:2.5 ratio from the line to each grid winding. That would be more than enough.

What is not clear is how to resolve the differences between driving 810's, which require much bias and less grid current, and driving 3-500's, which require minimal bias and 250mA average grid current (full power), 500mA peak grid current when pushed.

For one thing, the voltage to the plates of the 6080 would not need to be 400V. 100V peak is needed, so 200V is enough there to reserve a 100V drop on the 6080.

On the cathode side of the 6080, it looks at first like it would not need a negative 200V supply because the 3-500Z only needs -10-20V bias at the most. But something is needed for the 6080's bias, an also to drive the other 3-500's grid to negative 100V, so a supply of some kind needs to be there to keep the idle 3-500 from conducting while its plate voltage is highly positive, and the other one is conducting.

I was thinking about floating the 6080 bias supply, which is very small, referenced to the -15V @ 3A heavily bled bias supply now used with the 3-500's. This would give a negative bias voltage for the 3-500's that wouldn't go anywhere, and avoid a high voltage negative supply having to make lots of current. The alternative looks like having the 6080's pass a lot more current through their cathode resistors there in order to have a 15V bias instead of that required by an 810, and heat them resistors up. At peak grid voltage of +100V, the 3-500Z grid, I think, looks like about a 200 Ohm resistor, so those huge 15K resistors there are not right?

Anyway the diagram hopefully makes some sense. I tried to figure out the amount of feedback. There is no voltage gain in the 6080, but the 3-500 has a voltage gain of 25 in the circuit, so if I send just over 1% of the output voltage from the 3-500Z plate back to the front, it might be about 55V peak to peak and that would make a difference against the 200V peak to peak drive signal to its grid.

I know keeping the transformer was spoken against, but it is outside the feedback and is convenient. Does that make sense?

[WD5JKO]

Patrick,

    That circuit has possibilities!

Some things came into my mind as I read your post:

** What about using the 6336 instead of two 6080/6AS7?

** That negative feedback...at the 100K pot wiper, is the source impedance low enough to drive the secondary of the driver transformer? Or do you need something like a 6SN7 dual cathode follower stage there?

Since you are going to use those big low Mu triodes as a driver, consider a modification of the SEPP (single ended push pull) circuit to drive each 3-500Z tube. At the following link, refer to Figure 6:

http://www.audiodesignguide.com/otl/aria.html

This adds a phase inverter to each 6080, but more importantly, that passive pull down resistor on your cathode follower is replaced with an active element.

Good Stuff Pat..

Edit: The crisscrossed NFB loop you added appears to be positive feedback...

Jim
Wd5JKO

[K1JJ]

Patrick,

Looks like a plan.

I like Jim and Steve's suggestions -  and want to add something:  If you got rid of the input audio transformer and used a splitter tube circuit, you could add the negative feedback thru the splitter.  Then tap the NFB off the mod transformer secondary.

The point is, right now you don't have the modulation transformer in the NFB loop. If you eliminate the driver transformer, you can then include the secondary of the mod xfmr in the loop. (Only one transformer in a NFB loop is usually practical.)

In my 4X1 tetrode-connected modulators, I take the NFB off the mod xfmr sec and it certainly makes a difference considering how much phase distortion a mod xfmr can add to the circuit when unchecked.  I am able to dial in about 10 dB of NFB despite the mod xfmr in line.  

I haven't thought this idea completely thru for your circuit, but now's the time to consider changes if you do...  

T