Texas Star DX500 and high SWR

[The Mad Scientist]

Excellent choice! If the first "Tech" told you the trouble was a 3rd harmonic only 12db down from the fundamental when your Bencher TVI filter had no impact at all......run like hell to the next tech in line. That one was clueless and may not even know how to operate a spectrum analyzer. When an honest tech hits a problem he can't figure out, he'll tell you he's stumped. When a "fly by night" guy gets stumped, he'll baffle you with BS until you're just as confused as him. For instance, claiming a self oscillation is a 3rd harmonic.

That is what I got out of it as well. Something did not sound right.

I had similar issues with the one that I had which I had cleaned up quite a bit with the rework of the bias devider network and reworking the amplifier grounds. I do beleive going to the regulated bias source would help a bit better. I will see what effect this has as soon as I am able to get around to changing it over.

 

[The Mad Scientist]

I think you are right, Eric. Just need to get one of the amps off to them so they can look at it.

73,
Brett

Ok Brett.

However if they give you that same aggravating general blanket statement "I cannot duplicate the problem" along with **"I have bever heard of this before"**as they seem to want to tell everyone that sends them these units for repair for this issue.....Then send it to me and I WILL nail it down.

I do think you would be better off having a streight 4 built as you were telling me that you were interested in doing.

Also part of the reason they aren't seeing some of these issues is because they do their testing into a dummy load and not into an actuall antenna.


The issues with these units have been going on for quite awhile and in my opinion they have had allot of time to come up with some sort of solution. I do not think that they are genuinely interested in trying to remedy it because to them it functions good enough for what the amplifier is going to be used for and dosen't warrent the time and expense to justify any type of in depth diagnosis. This was my experence with trying to work with them on this issue.


There is allot of room for improvement....... Period!!


They are not known for their customer service and going above and beyond that they used to be known for. I no longer find these units as reliable as they once were. I will eventually build myself a streight 4 and dump the Texas Star which I sould have done years ago.

I have not spoken to Mike in quite sometime and I am wondering if he has seen more of these units come in for similar issues and if he has come up with a remedy for it?


your opinion may vary.

73's

 

[Shockwave]

That 10 uh choke on the output and how it connects to the front panel switches, DC line, key light, RF meter and bias feed are completely ridiculous! Now that I've finally seen a schematic of the amp http://www.cbtricks.com/Amp/txstar/dx400_dx500/graphics/ts_dx400-dx500_sch.pdf I can tell you why bias, choke and oscillation issues have plagued the Texas Star line forever.

The answer is so easy I'm amazed the guys at Texas Star would work so hard to cut one corner so short. All of these problems are the result of them using the wrong relay for the application!!! Someone had a million double pole double throw relays and used them when they needed triple pole double throw relays. With only two poles on the relay, they could only switch the RF input and output circuits through the relay and were left in the dark with figuring out how to turn the bias and meter light on in transmit.

Someone had the not so bright idea that they could create a return path for DC through the two chokes and the output RF contact of the relay in order to skimp on the 3P2D relay. Probably one of the worst ideas I've seen implemented with a linear amp ever. When you consider the chokes would have to drop all of the RF on the output of the amp before they could feed the bias circuit with clean DC, no wonder this design oscillates.

I would get a small single pole relay and tie its coil across the original so it keys at the same time. Lift L2 from the side of the board where it connects to R30 and the meter circuit. Connect this lifted end of L2 through the relay contact so it gets 12 volts when the relay is keyed. You should also remove L1 from the circuit after adding a bias relay since L1 can load down the RF output if its value drifts from excessive drive. I'll bet most peoples problems go away with this simple fix. Now there will be a million times less RF on the DC that feeds the bias circuit. Shame on Texas Star for trying to save a dollar.

This also explains why Texas Star had to use the highest level of negative feedback in any amp to even get it close to stable. No other amp had to use 39 ohm 4 watt resistors with .15 uf caps for negative feedback. One design mistake often leads to a bunch of Band Aids that never address the root of the problem.

 

[The Mad Scientist]

Well I'll be damned. All this time trying to get THEIR design to work and all the while all that needed to be done is change part of it. :headbang

Going to change that streight away in my own unit and see what pans out.

Good catch

 

[Mudfoot]

Subscribed and watching this most informative post.

 

[2RT307]

Excellent choice! If the first "Tech" told you the trouble was a 3rd harmonic only 12db down from the fundamental when your Bencher TVI filter had no impact at all......run like hell to the next tech in line. That one was clueless and may not even know how to operate a spectrum analyzer. When an honest tech hits a problem he can't figure out, he'll tell you he's stumped. When a "fly by night" guy gets stumped, he'll baffle you with BS until you're just as confused as him. For instance, claiming a self oscillation is a 3rd harmonic.

Shockwave, this tech honestly did just get his spectrum analyzer about a week ago. It is an ATTEN model, but I don't remember the number. So, Iam sure that he is still learning how to use it. This guy has always been honest and truthful with me. Never, ever charges a bench fee if he can't fix equipment. He spent about 3 hours playing around with my amp trying to figure it out and make it work for me.

We discussed my using a TVI filter, and I told him that when I previously put the bencher in line, it didn't have any effect. He told me that didn't make sense, and I agreed. It was at that point that I thought maybe my jumper could have been bad, so I bought a new one while I was there. We were still discussing that this was just a band-aid fix, and that the amp would still have the harmonic issue. I will try the filter again... hopefully tomorrow if I have time to throw everything back in the truck.

73,
RT307

 

[2RT307]

That 10 uh choke on the output and how it connects to the front panel switches, DC line, key light, RF meter and bias feed are completely ridiculous! Now that I've finally seen a schematic of the amp http://www.cbtricks.com/Amp/txstar/dx400_dx500/graphics/ts_dx400-dx500_sch.pdf I can tell you why bias, choke and oscillation issues have plagued the Texas Star line forever.

The answer is so easy I'm amazed the guys at Texas Star would work so hard to cut one corner so short. All of these problems are the result of them using the wrong relay for the application!!! Someone had a million double pole double throw relays and used them when they needed triple pole double throw relays. With only two poles on the relay, they could only switch the RF input and output circuits through the relay and were left in the dark with figuring out how to turn the bias and meter light on in transmit.

Someone had the not so bright idea that they could create a return path for DC through the two chokes and the output RF contact of the relay in order to skimp on the 3P2D relay. Probably one of the worst ideas I've seen implemented with a linear amp ever. When you consider the chokes would have to drop all of the RF on the output of the amp before they could feed the bias circuit with clean DC, no wonder this design oscillates.

I would get a small single pole relay and tie its coil across the original so it keys at the same time. Lift L2 from the side of the board where it connects to R30 and the meter circuit. Connect this lifted end of L2 through the relay contact so it gets 12 volts when the relay is keyed. I'll bet most peoples problems go away with this simple fix. Now there will be a million times less RF on the DC that feeds the bias circuit. Shame on Texas Star for trying to save a dollar.

This also explains why Texas Star had to use the highest level of negative feedback in any amp to even get it close to stable. No other amp had to use 39 ohm 4 watt resistors with .15 uf caps for negative feedback. One design mistake often leads to a bunch of Band Aids that never address the root of the problem.

Thanks for following up on this, Shockwave. I am hoping this fix cures the amp... probably will wait for Eric to test it on his! lol btw Eric, sorry I didn't call you back today... was very busy with the family at a local school carnival.

73,
RT307

 

[Shockwave]

Shockwave, this tech honestly did just get his spectrum analyzer about a week ago. It is an ATTEN model, but I don't remember the number. So, Iam sure that he is still learning how to use it. This guy has always been honest and truthful with me. Never, ever charges a bench fee if he can't fix equipment. He spent about 3 hours playing around with my amp trying to figure it out and make it work for me.

We discussed my using a TVI filter, and I told him that when I previously put the bencher in line, it didn't have any effect. He told me that didn't make sense, and I agreed. It was at that point that I thought maybe my jumper could have been bad, so I bought a new one while I was there. We were still discussing that this was just a band-aid fix, and that the amp would still have the harmonic issue. I will try the filter again... hopefully tomorrow if I have time to throw everything back in the truck.

73,
RT307

That being the case, he is probably making the common mistake of overdriving the spectrum analyzer. If the RF sampler driving the scope does not provide enough attenuation, it will cause the spectrum analyzer to exaggerate harmonics. Since we are dealing with a push-pull RF amplifier that has no filtering, we should expect the third harmonic to look worse than the second harmonic but 12db down is a lot. Push-pull amps cancel second harmonics well but do nothing to address the third.

The tech should also learn the difference between spotting the symptoms of harmonic problems and the difference between this and self oscillations within the amp. The oscillation problem you have is close to the fundamental frequency and below the 30 Mhz cutoff frequency of the Bencher filter. The filter will not help. The tech should be looking for spurs on either side of the carrier with the spectrum analyzer because the oscillation is within a few Mhz. of this. Keep in mind it still may only be triggered when operating into a resonant load.

 

[Mudfoot]

I'm sure your tech has honorable intentions, as does your amp builder friend.

Only thing I see that they might want to do, is start participating in online forums. I read where you mentioned that they don't do the forum thing yet.

The Internet is sure full of nut jobs, but doggone it, there sure are some downright intelligent dudes and dudettes.

 

[Shockwave]

Well I'll be damned. All this time trying to get THEIR design to work and all the while all that needed to be done is change part of it. :headbang

Going to change that streight away in my own unit and see what pans out.

Good catch

While I think isolating the bias feed from the output RF is going to stop all of the oscillation problems there are still other issues that could effect performance. We've always known the bias circuit sucks because it had no regulation or thermal feedback. My cure goes after the root of the RF feedback problem that causes the oscillation with a high SWR and does nothing to improve the basic bias circuit.

I would at least replace the 1.8 ohm resistors with a silicon diode that has a similar forward bias voltage drop as the final transistors. Mount these diodes over the closest final transistor cap with thermal compound to offer some degree of thermal tracking. Check that the value of the 25 ohm resistors is providing the correct amount of bias current to the finals.

Do this by inserting a current meter inline with each of the DC lines feeding the output transformers. Each pair of output transistors should draw about 200 ma when only the bias relay is keyed and no RF applied. If you find the current is too high, you can increase the value of the 25 ohm resistor or better yet, install an adjustable 3 terminal regulator to lower the voltage and further stabilize it. Bypass the regulator with .01 caps on the input and output.

Also remember they used very low value feedback resistors with very high value caps. Once all of the RF feedback through the bias circuit is removed, the additional negative feedback loop may provide too much feedback. It might improve gain and stabilization by changing these values to the more common 100 ohm 2 watt and .001 to .01 uf caps. You should also remove L1 from the circuit after adding a bias relay since it will load the RF output down if its value drifts from excessive drive.

 

[Robb]

OK, I'll bite.

So this poor choice of parts to make this amp work should be revised. If not by the mfr, then the owner or a shop.

But this amp that started this thread was working until the relay took a dump - right? Would it be easier for him to just replace the relay than to upgrade it to the more proper way you pointed out?

Not saying that the engineer did it right in the first place as you pointed out. They built it with cheaper parts to keep the cost down. Should he just replace the part/band-aid it?

 

[Shockwave]

OK, I'll bite.

So this poor choice of parts to make this amp work should be revised. If not by the mfr, then the owner or a shop.

But this amp that started this thread was working until the relay took a dump - right? Would it be easier for him to just replace the relay than to upgrade it to the more proper way you pointed out?

Not saying that the engineer did it right in the first place as you pointed out. They built it with cheaper parts to keep the cost down. Should he just replace the part/band-aid it?

The biggest problem here is that the existing relay contacts are being used to power the bias circuit. That requires the use of two chokes that attempt to isolate the DC power from the RF output that is switched by the same relay contact. The problem occurs over time because the 10 uh chokes are constantly stressed having to drop the entire output of the amp across the choke. This causes the choke to drift in value and loose some of its impedance to the RF.

The simple fix of just adding another small relay to apply the 12 volts to L2 in transmit along with the removal of L1 should make the amp run unconditionally stable. Fixing all the SWR issues too. Swapping out the 1.8 ohm resistors with forward biased diodes mounted on top of the finals is an extra step that would offer some regulation and thermal tracking.

Using the original voltage divider with the 1.8 ohm and 25 ohm resistors allows the forward bias current to drift significantly. Some have even suggested under full load the amp may drop into class C. Remember that the cars electrical system can easily drift plus or minus 10% of rated voltage. Combine this with the voltage drop on the power wires feeding the amp and you could double that to 20% in poor cases.

Keep in mind that a simple resistive voltage divider will pass this same percentage of voltage regulation right on to the load or bias circuit in this case. Any idea what happens to the idling collector current on the finals if the base voltage drifts 20% from .6 volts? If it drops to .5 volts, the transistors could be cutoff and only forward biased from the RF drive. At .7 volts of bias they could be close to saturation and this is why a stock Texas Star always blows up when you increase the supply voltage.

 

[Shockwave]

It's funny how much times have changed since the 1980's when these Texas Star designs first came out. The few I would see from time to time always had the 10 uh chokes burned out and would oscillate. There was no internet or published schematics for this series then. It wasn't worth my time to pull the board and make my own schematic of the double sided PCB.

That being the case, two things always baffled me. Why in the world did they use chokes of such critical value when no other amp used them? Second, how in the world did they get the meter light and bias to turn on in TX with only two poles on the relay? The picture of that schematic I found tonight was worth a thousand words to say the least. Turns out the two questions I always had may have led right to figuring out the problems we always see.

Dentron was another manufacturer that was infamous for using double pole relays on amps that were biased. To save a buck on the low end amps, they just left the bias on all the time. That would cost you a set of tubes fairly quick but at least it didn't oscillate.

 

[The Mad Scientist]

Thanks for following up on this, Shockwave. I am hoping this fix cures the amp... probably will wait for Eric to test it on his! lol btw Eric, sorry I didn't call you back today... was very busy with the family at a local school carnival.

73,
RT307

And test it I will!

Thanks for the insight Shockwave, I appreciate it much. I think that I'll leave the bias devider network for now untill I see what effect of seperating the dc will have. And since the 25 ohm sandbar resistors are unreliable as they do drift upwards over time reducing the bias voltage so eventually the devider network will go in favor of the regulated feed.


As far as the idle current I am not to keen on setting them at 100 ma. each. I am more comfortable having them set at around 60 for a total of 120 +or- 2 to 3% or so due to the extra heat load.

What would you recommend as far as value for the foward bias tracking diode?

In anycase, Perhaps ther IS a light at the end of the tunnel for these units after all.

No problem Brett, Drop me a line when you can.

73's
M.S.

 

[Shockwave]

And test it I will!

Thanks for the insight Shockwave, I appreciate it much. I think that I'll leave the bias devider network for now untill I see what effect of seperating the dc will have. And since the 25 ohm sandbar resistors are unreliable as they do drift upwards over time reducing the bias voltage so eventually the devider network will go in favor of the regulated feed.


As far as the idle current I am not to keen on setting them at 100 ma. each. I am more comfortable having them set at around 60 for a total of 120 +or- 2 to 3% or so due to the extra heat load.

What would you recommend as far as value for the foward bias tracking diode?

In anycase, Perhaps ther IS a light at the end of the tunnel for these units after all.

No problem Brett, Drop me a line when you can.

73's
M.S.

The 100 ma idling current is what the data sheet calls for on a 2SC2879 in class AB. Less will improve efficiency and reduce heat but you'd have to experiment to see how low you can go before creating objectionable distortion.

I usually used standard silicon rectifier diodes like the 1N4004 series. You just want one that has a forward voltage drop similar to the base-emitter junction of the final transistors.