NEW Project - 6146B Plate modulated by Hammond 1628SEA and 85W SS Amp

[W1ITT]

Tom....
Last year I bought a "digital soldering station" thing on Ebay for less than $100.  As with most of that ilk it's not there now, but others of the same type are.  It has a small hot air gun with controlled air temperature that you can dial up to a specific temperature.  This is sometimes referred to as "reflow" and it how the chip was soldered on in the first place.  You can use it to unsolder the fake chip and replace it with a genuine article chip, if your hands are still steady and your eyes are still fairly sharp.  And you will most likely do it without damage to board or components as all this stuff is designed for this treatment.
My unit also has a conventional small soldering iron on the other side that I can dial up to temperature and be soldering SMD chip capacitors and resistors within a minute,  It may not be an American Beauty, but it's still a beaut.
73 de Norm W1ITT

[WD5JKO]

Hi all, I once ordered some big three phase rectifier modules from a China based Ebay seller. I get an email from they seller, in translated english,
basically asking me the exact part number to brand on the side of the module! I knew this was bad, but they came in, looked perfect, and worked fine, over a dozen of them!!

As to the situation in hand here, a 4 or 8 ohm speaker has several ohms DC resistance in the voice coil. A slight DC offset from the amplifier will then only be a minor issue. Conversely, driving the 4 or 8 ohm winding on a transformer will be a huge problem if there is a DC offset, or if there is an asymmetry in the AC waveform. The winding resistance could be 0.1 ohm, and with a 1V DC offset would cause 10 amperes to flow! Might try two big electrolytics in series - + + - to form a poor mans AC capacitor. Something like 4700uf or thereabouts x2 at 25v would be a good test. Put the two series caps going between the AMP and the Transformer primary.

Jim
Wd5JKO

[W1RKW]

Tom,
to desolder the TDA, cut the long leads first close to the body of the IC then fold back the IC to get to the short leads then cut those as close to the body of the TDA as well.  Clamp the board down and extract the leads one by one by heating them up and pull out with needle nose pliers or tweezers. Tinning them a bit might help with heat transfer.  Clean up the holes with solder wick.  that's about the fastest and cleanest way to do it with basic tools with minimal board damage.

EDIT:  I looked at the link on counterfeit  parts.  Holy cow is all I can say.  I watched this youtube video:  https://www.youtube.com/watch?v=1IlTckCE8CM

[K1JJ]

Very interesting on all!

I gave it another try and this time I was able to get it working on the bench.  The rig was off.  I'm not sure what is different out of the rig. I just extended the leads with clip leads. I could smoke an 8 ohm 20 watt resistor with a clean sine wave.  But after about 5 seconds it folds back and looks crappy again.

I tried it with the rig running and Hammond transformer.  At about 15 watts carrier, I can modulate it up to about 250 watts cleanly.  But after 5 seconds it goes away and folds back again. Maybe I am seeing the effects of thermal protection shutdown, I dunno.  I am using a smaller heatsink on the bench and it gets warm quickly and works best and longest after a cooling rest.

I tried using shielded cables, big ferrite beads on all leads but it still folds back. There are also some parasitics mixed in when it happens.  BUT for the first 5 seconds the chip seems to be performing well.

Now thinking about ordering a soldering station as Norm described and ordering a REAL chip from DigiKey as suggested by Ed.    I've been thru this stuff before with the SS MOSFET audio driver and finally got everything working. Mostly layout and RFI getting in.

I might hook up the 150W Dayton analog amp that I use with the 4D32 and Hammond big 1643SE to prove out this rig.  It it works well with that outboard amp, then I can have confidence it can work eventually with the TDA7294 down the road.

Jim I already tried the back to back caps and no difference.  The module appears to like the 16 ohm best. It is most stable there.  The whole lash up appears to have big potential if I can tame it down.

Bob: FB on the chip removal procedure. Tnx.

* Some progress. I think it is thermal shutdown after 5 seconds.  I let it cool off and the waveform was clean for 15 seconds before the heatsink got hot and folded.  I'll return it to the inside rig on the bigger heatsink and try again.  It will handle a 65 watt carrier quite well with huge peaks when it is in a good mood..  Lots of things going on in there.   I may need a big shielding job  via an alum box around it.   It IS sensitive to higher RF power for sure.   At 36 dB gain, it drives easily with my station one volt audio source.

** Put the TDA back on its heatsink into rig and fired it up.  The triangle wave looks big and beautiful for the first 10 seconds and then starts to fold back.  The chip is staying cool on the heatsink with the fan blowing.  So something else is changing and folding it back. The signal going in is FB.  I will try a 16 ohm resistive load on the output without the rig running and go from there.  It may end up I buy a new REAL chip and replace it.


T

[K1JJ]

Update 3-24-2020:

There is definitely something wrong with the TDA module.  I replaced it with my Dayton 150W analog audio amp.  The rig ran perfectly with clean triangle waves and peak audio power. There was no limit for time.

So the TDA module wants to fold back after 5-10 seconds. The more audio power I draw, the quicker it drops out.  Maybe a bad critical part inside that is heating up?  I've seen this before.... runs good then gets flaky and jumps around like James Brown and dies. Over and over.

Anyway, I'm on the fence as to whether to get a new solder station and new chip from DigiKey -- or return this TDA to the eBay vendor and buy two more, one as a backup.  What keeps me hanging on is that it works perfectly for 5-10 seconds...   BUT, the circuit board itself may have a bad part which would be another waste of time if I replace the chip.

T

[W1ITT]

Tom... 
I just looked at the TDA7294 in a few examples on Ebay.  I think the hot air soldering station would not be the tool for this job.  They are meant for little tiny SMD devices and the ICs with pins so small that you can hardly see between them.  It's a fun tool, but I think it's for smaller stuff than the TDA, if what I'm looking at is the same thing.  I think I'd just suck it up and buy a couple whole replacement amplifiers and hope you can get one that was made on Tuesday.  If you are just honor bound to replace the chip itself, I'd do what someone else suggested and snip the leads then unsolder the debris and clean with a spring loaded solder sucker tool.  I think my old one was made by Ungar and still works.  But the hot air station just doesn't look like the man for the job.
73 de Norm W1ITT

[K1JJ]

OK, Norm, TNX.

Some progress: I probably won't need the precise soldering station anyway.  After a lot of testing and trying various stuff, I've found that the TDA works OK, but it does NOT like the Hammond transformer.

When run alone into a 8 ohm resistor it is clean up to 90 watts, no problems at all. But as soon as I connect the output to the 8 ohm winding it gets squirrely as I have described.  I tried the back-to-back capacitors to no avail.  (2  10,000 uF @ 50V)  The best tap is 16 ohms and the worst is 4 ohms FWIW..

Even if the TDA is connected to the 8 ohm resistor load and the transmitter is off, when the clip lead touches the Hammond primary winding it folds back.  This tells me it is not RF getting into anything, just that the TDA does not like an inductive load.  Again, to back this up, the Dayton 150 watt amplifier works perfectly at full strap. Any further ideas?

I want to use the Hammond and the general rig as wired. But unless I find a completely different audio module, I may pick up another Dayton 150 as an outboard SS amp.  

There is a possible advantage... it will leave room when the 35-0-35V TDA power transformer is taken out and the muffin fan moved closer to the tubes. I could add a strapping mod reactor and coupling cap to take the DC off the Hammond. Right now I am drawing about 200 mA+ and the Hammond is rated at only 130 mA DC current. I think this would give me more audio capability to modulate those three tubes better. Also, the DC would be off the xfmr for less breakdown arcing risk.

I already tested the rig with the Dayton amp and know it works FB. (even has 150W bridge mode into 8 ohms) The added room for the choke and additional headroom might work out.

So that's it for now.  This is what happens when we try untested ideas.  I might build the TDA and transformer into a nice station amplifier. I presently use a small Chinese 25W box - and the wall speakers wud like something like this. But I'm still open to ideas before I give up on the TDA....

T

Here is the Dayton 150... a very nice amplifier for about $140 used on eBay: (2 pages of data)

https://www.manualsdir.com/manuals/294570/dayton-audio-apa150-150w-power-amplifier.html

[WD5JKO]

Hi Tom,

   Take a look at the App Notes for the old LM12 power op amp, specifically on pages 8-10. See PDF attachment.

The clamp diodes from the output to each power buss could be important with your TDA, and easy to add.

Also, any audio transformer has a self resonant frequency, often up in the tens of Khz range. If the TDA has enough gain up there, it may oscillate at that frequency.

A possible fix is to insert a series R-C shunted across the transformer primary such that R is somewhere between 4-16 ohms and the reactance of C is low at the transformer resonant frequency, might try 0.5uf and 10 ohms.

BTW, I have two NOS LM12CLK devices dated 8849. Yours if you want them.

https://www.electroschematics.com/lm12-150w-audio-amplifier/

Jim
wd5JKO

[KD1SH]

I work in an R&D lab where I do lots of precision soldering. Pretty much the only thing I use the hot air pencil for is removing QFN and BGA and similar type packages where there are absolutely no exposed terminals - all solder connections are hidden under the body of the component. Otherwise, it's solder-wick for most SMD stuff with leads, hot-tweezers for SMD resistors and caps and such, and the vacuum solder-sucker for through-hole parts.
A good rule-of-thumb for soldering iron tips, whether soldering or de-soldering, though it might seem counter-intuitive: use the largest iron tip that will contact the leads of the part without incidental contact with any other parts. The larger tip has more mass and will transfer heat to the leads more quickly than a smaller tip will, thus requiring less contact time. The smaller tip will require longer contact and could result in thermal damage to the board material and the part. Faster is better.
And, while I agree with the previous advice to cut the leads, be very careful - a wire clipper cuts not by removing material, like a file or a saw, but rather by displacing material (it's basically a wedge), which means that it has the effect of momentarily expanding the lead's length, and can result in mechanical stress tearing foil traces from the board.

Tom... 
I just looked at the TDA7294 in a few examples on Ebay.  I think the hot air soldering station would not be the tool for this job.  They are meant for little tiny SMD devices and the ICs with pins so small that you can hardly see between them.  It's a fun tool, but I think it's for smaller stuff than the TDA, if what I'm looking at is the same thing.  I think I'd just suck it up and buy a couple whole replacement amplifiers and hope you can get one that was made on Tuesday.  If you are just honor bound to replace the chip itself, I'd do what someone else suggested and snip the leads then unsolder the debris and clean with a spring loaded solder sucker tool.  I think my old one was made by Ungar and still works.  But the hot air station just doesn't look like the man for the job.
73 de Norm W1ITT

[K1JJ]

Jim, as usual you appear to have a good solution.  I also got the schematic below from Jeff / W2NBC and hooked it up.  I tried 8 uF and 5 ohms and the TDA 7294 is now stable!  I'll try to trim the values back. But the resistor gets just luke warm and appears not to be hogging too much power for now.

So far so good. I'm hoping to optimize things and see how many tubes are best, best HV level, screen dropping and grid leak values, etc.

I'll keep ya informed.

Tnx for all the suggestions, guys!

T

[K1JJ]

3-25-2020 UPDATE:  Random thoughts...

After some refinements and tests, all it took was a 2 ohm resistor in series with the 16 ohm Hammond winding to make everything stable.  Though, I think I will still need a cap across the resistor to keep voltage drop at bay.  I'll tinker with it later.

I also found that ONE 6146B final was the optimum choice for the final.  The audio peak power limitation is as expected - the transformer and the TDA 7294.  This combination is ample enough to modulate a single tube at 40 watts output @ a very robust 150% positive, but any more carrier and there is saturation.  

That's OK cuz it was my original plan to build a Ranger III for QRP and to drive the 4-1000A linear. .  The frequency response is quite good and very clean. However, it starts rolling off at 5 KHz which is odd. Alone into an 8 ohm resistor, it sweeps flat over 40KHz.  I must look at the screen bypass caps.  I only use 500 pF in the plate circuit.  Why is it that every new rig has frequency response limitations until it gets worked over?

I see no signs of audio foldback or parasitics with the 2 ohm 10 Watt in place, though if removed from the circuit the problem reappears.  I can live with the small heat loss that is barely luke warm after a sine wave tone test of 1 minute.

BTW, the TDA-7294 shows barely any heat after an extended sinewave tone test of 100% modulation. This is a very efficient method of modulation. It really didn't need a muffin fan.  The single 6146B itself barely shows any color after a 50W carrier and 100% tone test. The 100 Hz triangle waveform looks FB and sharp.  I settled on about 650VDC plate voltage for now.

I feel good about the "Ranger level" 40 watts out with lots of headroom audio. Parts are way conservative.  There are plenty of rigs in the shack that will do much higher power levels. Commander is now the baby.  A FB low power rig of high efficiency.  The rig is as quiet as a church mouse with the muffin fan turned down.  Fantastic.

Still lots more testing and refining left to do.  I'm happy I didn't have to start over with the audio section...   I may try out the Hammond secondary spark gaps later.

Thanks to the guys in this thread -   and Jeff/ W2NBC and Frank/GFZ for their email suggestions.

T

[K1JJ]

UPDATE NOTES:

A few things...  I found the Bird wattmeter does NOT read audio tone peak power well.  I was watching the peak reading bird and thought it was rolling off at 5 KHz when in reality, I looked more closely at the scope and SDR analyzer and see we are flat up to about 20 KHz. And it even goes higher to 40 KHz with some roll off but still a clean sinewave.  It is near impossible to get that kind of performance with a standard tube plate modulator.  In the past I can remember saying some sss's and the Bird hardly moved. Now I see why.

Also, I ran the rig hard for a couple of hours with the muffin fan off. The TDA-7294 DOES get warm to the touch thru the heatsink but only a warm, not hot at all. I turned the muffin fan on and in a minute it was back to room temp.

I'm trying to find the right values for the resistor and cap for the 16 ohm input.  I want to get as much power out as possible but still keeping it stable.

I did a harmonic distortion test  and see that side products are down at least -35 dB.  This is good and shows no problems. I was able to sweep the audio sine wave up into the 20KHz and higher ranges and see the full signal dancing up the band. The bandwidth is quite impressive. I will use the DSP audio filter to be a good neighbor, of course.

So far nothing has crapped out or smoked since the rig was built. Just one 16 ohm fuse popped when abusing it. This is highly unusual and needs a full house investigation to figger out why..  

T

[KC2ZFA]

'JJ

you still have the series 2.7 ohm .1 uF across the output of the TDA module ?

Peter

[K1JJ]

'JJ

you still have the series 2.7 ohm .1 uF across the output of the TDA module ?

Peter

Hi Peter,

Yes, the 0.1 uF and 2.7 ohm are there on the board.  I tried padding it with different values, but no difference with the problem discussed below.

I've made some big changes since the last post.  After doing a closer inspection of the rig's performance, I found that Commando works very FB, BUT there are some small audio parasitics riding on the desired audio whenever I put a 60 Hz or lower tone thru.  The tone triggers the parasitics.  I had been fighting this problem in different forms for a few days and thought it was licked.   I tried everything I could think of and also talked with a couple of friends who know the score. I could have rolled the audio off below 60 Hz and been done with it, but decided I didn't want a band aid rig with a hidden problem.  This particular TDA-7294 module just can't hack the Hammond below 60 Hz.   It may be just this particular TDA with a fake chip, but I had to cut my time losses short and move on.

A more powerful and dependable solution: My good friend, Jeff, W2NBC offered to send me his spare Hammond 1642SE, the big MOFO 28 pounds, 55 Henry 300 mA beast.  I will be using it with an old Dayton 150 Watt audio amp I've used for years with Summer Breeze, the 4D32 rig.  I put the three 6146s back in and tuned up Commando.   I cut out and installed a slick AC power socket outlet on the back of Commando that will turn the Dayton on from Commando's front panel. I also added a -41 volt fixed grid bias supply and got rid of the screen clamper. The clamper worked OK, but the three 6146s were a little too much for it load-wise.

So I'll be experimenting to see what is the best tube complement.  (1,2 or 3 tubes)   The Hammond is good for 300 mA of DC current so would be a good match for two tubes anyway.  

The space that was freed up by taking out the +/- 35V power transformer, the fan and the heatsink will easily accept the Hammond 1642SE onboard. The Dayton amp will sit by itself out of sight behind Commando remote controlled.   I know this lashup will work FB since I tried a similar version already - and is an improvement to making it a super Ranger if not a good match for a Valiant power.

I relocated the muffin fan on the inside front panel closer to the tubes, directing air downwards. At 30% spin rate it keeps the finals in a good state.

I'll take some new pics when the new Hammond is installed and working.

So bottom line is anyone building the original TDA version should run a bench test with the transformer hooked up to the TDA  and into a 5K resistive load.   Look for tiny buzzing parasitics below 60 Hz when drawing audio power via a tone below 60 Hz.   Be sure to try a TDA with a real chip.  I estimate the parasitics to be around 300 Hz or so with a pulse character to them.  I did not measure the frequency.


T

[Chuck...K1KW]

Yo Tom,

Why the TDA-7294?

You should be using the TDA-7293.

Check out the differences in the specs.

The '94 can not handle what the '93 will do into low impedance loads which is what you get when going lower in freq into a transformer.  Probably why the '94 craps out below 60 cps.  Also, the '93 has a higher voltage capability in addition to the much better load tolerance. 

I've got my '93 based board driving the 1628 into a dummy load at 60W cleanly down to 7 Hz.  That is with no DC current however so this isn't a good test. 

Getting ready to modify my Ranger with this set up.


-Chuck

[KC2ZFA]

Tom, thanks for the info. Good thread.

Peter

[K1JJ]

Yo Tom,

Why the TDA-7294?

You should be using the TDA-7293.

Check out the differences in the specs.

The '94 can not handle what the '93 will do into low impedance loads which is what you get when going lower in freq into a transformer.  Probably why the '94 craps out below 60 cps.  Also, the '93 has a higher voltage capability in addition to the much better load tolerance.  

I've got my '93 based board driving the 1628 into a dummy load at 60W cleanly down to 7 Hz.  That is with no DC current however so this isn't a good test.  

Getting ready to modify my Ranger with this set up.


-Chuck

Interesting info, Chuck, thanks!

If you were able to get your TDA7293 to work with the transformer and load connected as described, you should be OK.    7 Hz low end is amazing.   My TDA7294 took off even without the rig connected, no DC - it just required the transformer connection with or without a load to misbehave.

So what in the TDA7293 datasheet (below)  stands out that it would be a better choice for low frequency, low inductive loads?   I do like the higher 100V and 100W specs.  I may still modify a real Ranger later on if you get yours working well.   These little rigs are so much fun to play with.

https://www.st.com/resource/en/datasheet/tda7293.pdf

T

[Chuck...K1KW]

If you overlay the graphs for the '93 vs the '94, it's obvious on the 4 vs 8 ohm loads vs voltage.

Plus, I've got both the '93 and '94 boards when I bought these from epay.  I bought a bunch of them from the various chink vendors.  Tried them both, the '94 board had issues like you found and the '93 board had none and drove my 1628 just fine into the dummy load.  Interesting that the '93 and '94 board were the same except for the chip!  Different chink vendors...

Also, on one of the HiFi forums, there was a discussion about these two chips.  A Google search will help but I didn't save it plus you will find lots of interesting info vs the two.  Bottom line is that the '93 is much better and there are a lot of '94's that were chink labeled as '93's.  So if you have a '94 it's really a '94.  Try a real '93.   There is a you tube video of a guy comparing the two on a subwoffer that doesn't reveal anything...so ignore that.

[K1JJ]

Chuck,

OK on the 4 / 8 ohm loads vs: voltages datasheet comparison for the '93 vs: '94.      

Later on after I get the Dayton amp going, the overall '93 concept  still seems like a good way to go for me. More compact and all onboard. Also, I like the idea of bridging a pair to get 200W.  There's examples of that on YouTube too.  I just received the big 1642SE Hammond from Jeff today, so will be testing that with the 150W Dayton amp soon.  But down the road I will likely look into a pair of '93s since I will have more audio power capability to play with using the 1642SE and three 6146Bs.  We'll see how your Ranger mods work out first.

Thanks for the info and shedding some light on this audio parasitic problem.    

T

[K1JJ]

We're now in the testing mode.  The Hammond xfmr 1642SE is strapped on and working.  The Hammond got slightly warm as well as the 800V power transformer.  The Hammond popped my arbitrary 5A fuse in the 8 ohm line so I replaced it with 8A.  The spark gap was arcing at first so it's been widened. I can hear some talk back from the core with tones, but we'll see how it works on voice.  The rig must weigh 80 pounds now.  So far I'm very pleased with most performance parameters.

160, 75 and 40M are optimized for tuning and power out.  160M took  an additional coil and it needed some loading padding. But all bands put out 125 to 150W carrier at about 700V.  I can modulate the 125W carrier to at least 150% positive, just what I wanted. It will do a 10 Hz sinewave looking good. Check out the pic below of 20 Hz at 100% modulated.  It sweeps to 20 KHz with a FB waveform.  The Hammond is amazing.

I optimized the screen dropping resistor  to 20K and the grid leak to 5K plus -41V fixed protective bias.  Total of 140V screen voltage and -105 grid voltage, class C, carrier conditions.

The freq counter is working and uses a loop coil pickup near the tank coil. I tuned up on 20M by accident when I intended 40M and it let me know. That's a good check.

The band switching is fast, maybe 30 seconds.  The big rigs are very slow in comparison; too much rowing around town with the vac variables.  With the preset labels on the panel, it will be even faster. I want to get more active on 160 and 40M, so will be putting up a pair of stacked 40M dipoles at 130' and 65' fed in phase for USA coverage.  The ant on 160M is a dipole at 190' so we are looking forward to seeing how well 100W can do on AM.

The 6146Bs run gray plates, though I am working on plate efficiency right now.

There is rear chassis space on top if I want to put in another TDA7293 audio module and power transformer.

The HV step start and -41V fixed bias  (and dropping resistor) all work well together with no screen problems.

For the first time on the back panel I put in a proper AC cord receptacle and even a 120VAC socket to feed the Dayton amp. The Dayton is turned on by the Commando front panel switch.

Commando is turning into a handsome soldier if I may say so. I didn't take any pics of the bottom until the testing is over and I can tie wrap up the wires neatly.  All Teflon wire, of course.

The input L/C network gives about 1:1 for each band when tuned.

I still need to do some triangle testing and voice tests as well as key down torture tests. My philosophy is to beat the heck out of a rig during testing, looking for failure parts. Then back it way down to operate and usually nothing will fail for long periods. The only failures are human error mistakes.

I might even have Commando  on the air this weekend or early next week.  Ya never know.

More later -

T


First Pic: Commando actually transmitting a 125W carrier on 3885.00 into a dummyload.

2nd pic:  It's hard to appreciate how heavy and massive that Hammond is for a little rig. It's a 28 pound brute with 55 Henries of low end strap. Notice the wide washers used to support the chassis better top and bottom.  The fan has been relocated and does a great job cooing the 6146Bs.

Pic2:  That plate coupling cap is the size of a lemon. Only thing I had...     Fan on rubber feet - no noise.

[K1JJ]

Pic 1:  First time doing the AC cord right!

Pic 2:  20 Hz 100% modulated.

Pic 3:  Both coils used for 160M, so no tap.  Full big coil tap for 75M.     1/2 big coil tap for 40M.

[K1JJ]

Commando is nearing graduation...

Pic 1:  Close and personal with the trio running the show.

Pics 2 and 3: Probably the best photos of the bunch.  

Pic2: Notice the Bird wattmeter deflection on the right.

Pic 3:  Dayton 150 amp remotely powered  by AC socket.

[K1JJ]

Update  4-4-2020:

Some changes... back to the start again.  The rear chassis empty space has been filled in with the +/- 36V audio power transformer and soon to be TDA7293s.  The chassis and panel are now filled! No more room.

Here's why:  After successfully running Commando hard using the single 75W audio channel, I doubled the audio by bridging the two channels to make 150W. I did not like the performance when bridged.  

So, I decided to go back to the TDA7293, but this time I ordered a DIY kit board for $9 that has TWO chips and claims  140W.  This shud be enuff to hit the audio hard but still not blow out the Hammond if there is a surge.

Also, because of the fake chip issue, IE, boards with pirated TDA7294 chips, I ordered a few new REAL TDA7293 chips from Mouser for only $5.25 each. I'll populate the kit boards with these REAL ones right from the start;  fake chip problem solved.

I always did want a self-contained rig anyway.  The downside is I gotta wait two weeks for the China shipment.... a slow boat from China.

To see the TDA7293 eBay $8 DIY 140W kit using two chips in parallel, search on eBay:  (The link is way too long... why do they do that?)

"TDA7293 Power Amplifier Board Mono Audio Module Kit DIY 140W"

More to come.

T

[Opcom]

Pic2:  That plate coupling cap is the size of a lemon. Only thing I had...     Fan on rubber feet - no noise.

I was just going to say that cap looks really cool there.

I see the fan mounting. Those mounts remind me of the rubber caps that go on the floor-end of a cane. I also like the back-set screen meter. Is that due to too much high voltage?

[K1JJ]

UPDATE  4/6/2020:

Rick, W8KHK has given me some great information via emails based on his vast experience with solid state amplifiers.   My particular problem is audio talkback from mic to transformer.  The internal thermal resistance is a factor that can put the device on a point of no return. Rick has taken a close look at the TDA7293  and has made some suggestions. I am still working on my rig as planned and hope to have the new TDA boards here within 2 weeks.

Pat: Thanks for the comments. The screen meter is offset from the panel for safety. When using a screen dropping resistor, there are scenarios where 300VDC+ can appear on the screen, thus the meter was floated. The other meters are at low potential.


I received Rick's permission to post the info below. If you are planning to do something with these types of devices, this is a worthwhile read:

---------------------------------------------------------------

Rick said:

I have been following the Commando progress, but have just stayed back on the sidelines, quietly cheering you on.   With lots of experience on both tube and solid state power amplifiers, I can say with confidence that heavy strife testing of tube equipment rarely results in a crap-out, but with the solid state devices it is very easy to reach a point of no return, especially with extended sine or triangle wave testing.  These amplifiers are intended for voice and music, but not constant banging with the 100 percent duty cycle testing.

The reason I bring this up is to help you prevent the same type of failure on the new dual TDA-7293 module.  If you look carefully at the spec sheet, under the section 4.2 High Efficiency, you will see some cautions.  I will post some of it here and highlight the critical phrases.......

You may download the specific datasheet here:  https://www.st.com/resource/en/datasheet/tda7293.pdf

Their high-efficiency amplifier is similar to the class H modulator typified by the build Steve/QIX did a few years ago.  Note that in the ST TDA schematic for high efficiency, they implement dual plus-minus 50 volt supplies for the peaks, and dual plus and minus 25 volt supplies for the average power, and employ external darlington pairs on the pos and neg rails to take on the excessive dissipation external to the chip modules.  

My concern is that you might reach a failure point very quickly if testing at max power using sine waves, if you are just using the TDA device without the external darlington pairs in your build.  You should be fine with voice duty cycle, but I would certainly tread lightly if doing sine or triangle testing at max power.

Hope this helps..

73, Rick


Here is the excerpt with highlights:

  High efficiency Constraints of implementing high power solutions are the power dissipation and the size of the power supply. These are both due to the low efficiency of conventional AB class amplifier approaches. The circuit below in Figure 8 is a high efficiency amplifier which can be adopted for both hi-fi and car-radio applications. The TDA7293 is a monolithic MOS power amplifier which can be operated with a 100-V supply (120 V with no signal applied) while delivering output currents up to ±6.5 A. This allows the use of this device as a very high-power amplifier (up to 180 W peak power with THD = 10% and RL = 4 Ω); the only drawback is the power dissipation, hardly manageable in the above power range. The typical junction-to-case thermal resistance of the TDA7293 is 1 °C/W (max = 1.5 °C/W). In worst case conditions, to avoid the chip temperature exceeding 150 °C the thermal resistance of the heatsink must be 0.038 °C/W (at a maximum ambient temperature of 50 °C). As the above value is practically unreachable, a high efficiency system is needed in those cases where the continuous average output power is higher than 50 to 60 W. The TDA7293 was designed to work also in a higher efficiency way. For this reason there are four power supply pins: two intended for the signal part and two for the power part. T1 and T2 are two power transistors that only operate when the output power reaches a certain threshold (for example, 20 W). If the output power increases, these transistors are switched on during the portion of the signal where more output voltage swing is needed, thus "bootstrapping" the power supply pins (13 and 15). The current generators formed by T4, T7, zener diodes Z1, Z2 and resistors R7, R8 define the minimum drop across the power MOS transistors of the TDA7293. L1, L2, L3 and the snubbers C9, R1 and C10, R2 stabilize the loops formed by the "bootstrap" circuits and the output stage of the TDA7293. By considering again a maximum average output power (music signal) of 20 W, in case of the high efficiency application, the thermal resistance value needed from the heatsink is 2.2 °C/W (with VS = ±50 V and RL = 8 Ω). All components (TDA7293 and power transistors T1 and T2) can be placed on a 1.5 °C/W heatsink, with the power darlingtons electrically insulated from the heatsink. Since the total power dissipation is less than that of a usual class AB amplifier, additional cost savings can be obtained while optimizing the power supply, even with a large heatsink.  



Just out of curiosity, I ordered a couple of the dual TDA-7293 modules like yours, and a few of the authentic chips from Digi-key.  Shipping was less than what Mouser wanted for a small order.

You will probably be up and running before I receive my package from across the pond, but the price was right, I was tempted, so HERE WE GO!

I am most interested in the THD spec, where they list it as around 10% at full output.  That seems kind of high, and I am hoping that the numbers are much better at lower sustained power levels.

I assume you are going to use some large storage filter caps external from the board, as the tiny two on the board are just local bypasses for high frequency impulses, and they will not provide any storage for low to mid freqs.  I think the earlier module required only an external power transformer, where this one needs a fully-filtered supply.  I do not think regulation is in order, so long as the caps are BIG and the transformer resistance is low.

I have a Sony power amplifier with a couple dead modules, but it has a husky power supply and massive heat sink.  So I was thinking this is an opportunity to replace the guts and have a decent amp for the surround sound in my home theatre, UNTIL I finish the tube amplifier with a quad of 6336 carbon anode bottles.  These are like 6AS7/6080s on steroids, double the dissipation, so a pair per channel should easily do 100 watts of clean triode power.  But in the meantime I can probably tolerate a solid-state amplifier, and listening tests will probably tell more than typical distortion measurements.  I am NOT one of those audioPhools, but I do understand that steady-state measurements do not tell the whole story when it comes to impulses and attack time, where ringing due to inverse feedback is very apparent with music, but undisclosed with a sine or triangle repetitive waveform.

I am of course interested in any thoughts you might have......



Tom, there is one other option you might consider to protect the TDA module during turn-on and turn-off surges....

The module has a standby input and a mute input.  I do not know if your printed circuit board exposes this circuitry with an interface terminal.  If it does, you may protect the device, and it is ok if both standby and mute are tied to the same input through RC circuits.

First, figure the time constant of the RC circuit interfacing the mute and standby input signals to the module.  The spec sheet recommends a separate 5 volt supply to control these inputs.  It may or may not be necessary to use a separate supply, as you may have another DC supply which could be regulated and isolated for this control function.

The idea is to make sure the standby (and optionally, the mute) control lines are not enabled until the power supply is fully up.  If there is not enough delay on the built-in RC circuit to accomplish this delay, it is easy to add a 7805 sourced from some other supply, followed by an R and C that will delay the enabling of the TDA until the main power supply is up.  Just milliseconds, but you really want the two pos and neg rails up and stable before enabling, such that an offset between the two is not passed on to the output transformer.

For elimination of surges during shutdown, you could take a main power switch with an extra contact, and use that to remove the standby signal from the module, connecting it to ground instead of the 5v supply.  This is best done by a diode to discharge the capacitor (AT THE MODULE) if there is an RC circuit for this function on the PC board. (Discharging via the exposed standby terminal on the PCB may discharge very slowly, due to a resistor in series with the capacitor.)  
This will ensure the module is disabled to standby status before the main filter caps start to discharge.  If you have an external R C circuit in addition to the one on-board, then BOTH capacitors should be discharged by separate diodes connected to the power switch, or, optionally, a relay that is triggered, (without delay at turnoff), by the power switch.

This technique will provide considerable protection to the module at turn-on and turn-off.  At other times, the caveat is to make sure your audio chain has adequate limiting and clipping, such that there are no impulses overdriving the amplifier.  Of course, audio connections to the module should be secure and not connected or disconnected while the module is enabled.  That little thing packs quite a punch for its size, and you need to help him make sure he does not punch himself!