Sheldon, I feel compelled to explain a few things.
Balanced or SE input is no different for the TPM if you adjust the gain of the TXD(6db) to account for half "input" swing. The output of the TXD will still be exactly the same. So you need not adjust anything on the TPM, as the load it sees and the load the TXD sees are identical either wat. There is no difference at all in the way they work accept that you will not get the same benefit of the common mode rejection at the TXD input which you would get going balanced input. That is literally all.
The TXD is current feedback, so you have to think in those terms.
Adding a SE/Bal converter will do you no good really, as the TXD already handles that marvelously, but adding a balanced source will help a lot.
Be careful about your comparisons, the BPA is a totally different beast from a TXO. The only really similar thing is the servo.
Good luck, and sorry if I left you befuddled.
Cheers!
Russ
Balanced or SE input is no different for the TPM if you adjust the gain of the TXD(6db) to account for half "input" swing. The output of the TXD will still be exactly the same. So you need not adjust anything on the TPM, as the load it sees and the load the TXD sees are identical either wat. There is no difference at all in the way they work accept that you will not get the same benefit of the common mode rejection at the TXD input which you would get going balanced input. That is literally all.
The TXD is current feedback, so you have to think in those terms.
Adding a SE/Bal converter will do you no good really, as the TXD already handles that marvelously, but adding a balanced source will help a lot.
Be careful about your comparisons, the BPA is a totally different beast from a TXO. The only really similar thing is the servo.
Good luck, and sorry if I left you befuddled.
Cheers!
Russ
Russ Thanks for the extra words it helps me to understand the 4131 a lot better.
By the way one thing that has made it tougher is I had to figure out from the BOM as you do not have the components identified
e.g. R1 1.1k, in some cases not all the pieces are listed in BOM .
Right now there is still no 22.k1 resistor in the TXD BOM. it is missing!
So there are little discrepancies from schematic to text to BOM
here and there.
There are no schematics for the LCPS or the TXPS I have read all of the material over and over and even used the pictures to help figure out what was what. TXD in particular.
I know you are working on the DAC which I am sure is quite complex.
The whole TX thing is fairly complex
Best regards
By the way one thing that has made it tougher is I had to figure out from the BOM as you do not have the components identified
e.g. R1 1.1k, in some cases not all the pieces are listed in BOM .
Right now there is still no 22.k1 resistor in the TXD BOM. it is missing!
So there are little discrepancies from schematic to text to BOM
here and there.
There are no schematics for the LCPS or the TXPS I have read all of the material over and over and even used the pictures to help figure out what was what. TXD in particular.
I know you are working on the DAC which I am sure is quite complex.
The whole TX thing is fairly complex
Best regards
SheldonD said:
Yes you are right, and we have tried to fill the voids with the documentation on the site, but I agree it is still lacking a bit. It is something I will personally strive harder to rectify. Thanks for helping me see it through a user's eyes. Sometimes I get "designer's goggles" syndrome.
I am working on more complete TXO documentation. And I am writing DAC documentation even tonight.
So I think you will see some improvement there.
Cheers!
Russ
Re: txd
RO is there to protect the TXD from capacitive loads. Because the TXD will not always be used to drive Twisted Power modules which are not capacitive loads I wanted to be sure there was a spot there for those resistors should they ever be needed.
For TXO you just use jumpers in place of RO1 and RO2.
For single ended input into a TXO you bring your input into +IN and connect -IN to GND. If you plan on using the TXO this way I would configure the amp for 20X gain.
Cheers!
Russ
tryonziess said:
RO is there to protect the TXD from capacitive loads. Because the TXD will not always be used to drive Twisted Power modules which are not capacitive loads I wanted to be sure there was a spot there for those resistors should they ever be needed.
For TXO you just use jumpers in place of RO1 and RO2.
For single ended input into a TXO you bring your input into +IN and connect -IN to GND. If you plan on using the TXO this way I would configure the amp for 20X gain.
Cheers!
Russ
Re: txo 8
You should have no trouble doing a TXO-8 but I can't imagine why you would need that much current.
I have successfully tested a TXO-6. I doubt a TXO-8 would behave any different. When I get time I may try it out myself. But the only application I can think of is 2ohm subwoofer.
Cheers!
Russ
tryonziess said:
You should have no trouble doing a TXO-8 but I can't imagine why you would need that much current.
I have successfully tested a TXO-6. I doubt a TXO-8 would behave any different. When I get time I may try it out myself. But the only application I can think of is 2ohm subwoofer.
Cheers!
Russ
txo 8
running 20 speakers in a line array necessitates series parallel wiring in excess. if i use a lower series count ie, increased current, maybe it would make individual speaker characteristics less noticeable. i had thought of buying some 72 inch bohlender ribbons which might need the extra juice. just fun toys, beats watching tv. anyway they would run cooler with more chips in parallel. thanks for the input. first amp almost ready to heat up will be buying another quite soon.
running 20 speakers in a line array necessitates series parallel wiring in excess. if i use a lower series count ie, increased current, maybe it would make individual speaker characteristics less noticeable. i had thought of buying some 72 inch bohlender ribbons which might need the extra juice. just fun toys, beats watching tv. anyway they would run cooler with more chips in parallel. thanks for the input. first amp almost ready to heat up will be buying another quite soon.
Import note about parallel TPMs TXO-4 and up
Last night I did some load testing of a TXO-6. When I built the TPMs I closely matched all of the feedback resistors (R1 and R2). That amp performs very well. What I wanted to test was the effect of gain imbalance between the power modules. So I intentionally mismatched one of the feedback resistors to get a little higher gain on that TPM.
What I found was that the module with the higher gain was supplying nearly double the current of the other modules.
The TPM includes an output current sense resistor (it actually works like local feedback). which helps a bit, but at .1R not really enough if the resistors are 1% off.
So if you plan on build a TXO-4, 6, 8 etc be sure you use .1% or better resistors (or resistors hand matched to that tolerance) for TPM resistors R1/R2 and possibly increase the size of the output resistor(R7) to .22R. Both of these measures together will insure that the TPMs distribute the load evenly.
We will be making appropriate changes to the BOM, and we will be in contact with anyone who got a TXO-4 to get you some replacement resistors.
Keep in mind, I listen to a TXO-4 with 1% resistors everywhere, but I hand matched them. This is fine. But I just wanted to make sure people knew the importance of gain matching to optimal performance. Without it one TPM will quickly reach thermal protection, which would sound, well icky...
Another note, this only applies to TXO-4,6,8 etc. TXO-2 because it is not running parallel power modules is not effected in the least.
Cheers!
Russ
Last night I did some load testing of a TXO-6. When I built the TPMs I closely matched all of the feedback resistors (R1 and R2). That amp performs very well. What I wanted to test was the effect of gain imbalance between the power modules. So I intentionally mismatched one of the feedback resistors to get a little higher gain on that TPM.
What I found was that the module with the higher gain was supplying nearly double the current of the other modules.
The TPM includes an output current sense resistor (it actually works like local feedback). which helps a bit, but at .1R not really enough if the resistors are 1% off.
So if you plan on build a TXO-4, 6, 8 etc be sure you use .1% or better resistors (or resistors hand matched to that tolerance) for TPM resistors R1/R2 and possibly increase the size of the output resistor(R7) to .22R. Both of these measures together will insure that the TPMs distribute the load evenly.
We will be making appropriate changes to the BOM, and we will be in contact with anyone who got a TXO-4 to get you some replacement resistors.
Keep in mind, I listen to a TXO-4 with 1% resistors everywhere, but I hand matched them. This is fine. But I just wanted to make sure people knew the importance of gain matching to optimal performance. Without it one TPM will quickly reach thermal protection, which would sound, well icky...
Another note, this only applies to TXO-4,6,8 etc. TXO-2 because it is not running parallel power modules is not effected in the least.
Cheers!
Russ
resistor match
was this test at or near clipping. change all of them to .1. also, if i mount the chips directly to the heatsink without insulation and insulate the heatsink from the chassis will the set of tpm on a particular sink react or stay neutral ie,distortion, hum, static. two separate sinks one + input the other - input in bridge. i read direct contact made for much better heat transfer.
was this test at or near clipping. change all of them to .1. also, if i mount the chips directly to the heatsink without insulation and insulate the heatsink from the chassis will the set of tpm on a particular sink react or stay neutral ie,distortion, hum, static. two separate sinks one + input the other - input in bridge. i read direct contact made for much better heat transfer.
Russ; interesting test.
How close did/could you match the 0.1 resistors. although this would be hard to do without resorting to a bridge.
I think these should need to be 0.1 % as well
The writeup on the LT1083 shows for parallel LT1083s
2 feet each of #18 wire to arrive at 0.015ohms.
Talk about matching!
How close did/could you match the 0.1 resistors. although this would be hard to do without resorting to a bridge.
I think these should need to be 0.1 % as well
The writeup on the LT1083 shows for parallel LT1083s
2 feet each of #18 wire to arrive at 0.015ohms.
Talk about matching!
Re: Re: Import note about parallel TPMs TXO-4 and up
I have a few hundred of each of those values (10K and 1K 1%) and it fairly easy to make a matching set.
I only matched them as well as my 6.5 digit DMM will allow. It is only their value relative to each other that is important.
Of course, the advantage of buying .1% resistors is that the tedious matching is eliminated.
I have to admit I am puzzled by your question, but I will try my best...
As far as my design goes It would not matter at all if the parts used are 4 (6 or 8) LM3886 or 2(3 or 4) LM4780 if you were being faithful to the TXO design. It would actually be pretty easy to do a TXO-8 using just 4 LM4780, but the thermal characteristics would not be as good as the single amp package. Still, it should work just fine. The circuit would be the same.
Cheers!
Russ
SheldonD said:
I have a few hundred of each of those values (10K and 1K 1%) and it fairly easy to make a matching set.
I only matched them as well as my 6.5 digit DMM will allow. It is only their value relative to each other that is important.
Of course, the advantage of buying .1% resistors is that the tedious matching is eliminated.
SheldonD said:
I have to admit I am puzzled by your question, but I will try my best...
As far as my design goes It would not matter at all if the parts used are 4 (6 or 8) LM3886 or 2(3 or 4) LM4780 if you were being faithful to the TXO design.
It would actually be pretty easy to do a TXO-8 using just 4 LM4780, but the thermal characteristics would not be as good as the single amp package. Still, it should work just fine. The circuit would be the same.Cheers!
Russ
Russ: I meant about matching the output load balancing resistors.
It is important here as well as the matching of RF and RG.
This is why I gave the LM1083 example. I f there is an imbalance here thn the different amps can be supplying different currents even if RF and RG were perfectly matched.
It is tough with the average meter to get 1 out of a thousand accuracy (0.1%) when the device under test is only 0.1 ohms.
Just matching to this would be difficult.
Your super meter does go along way, though.
It is important here as well as the matching of RF and RG.
This is why I gave the LM1083 example. I f there is an imbalance here thn the different amps can be supplying different currents even if RF and RG were perfectly matched.
It is tough with the average meter to get 1 out of a thousand accuracy (0.1%) when the device under test is only 0.1 ohms.
Just matching to this would be difficult.
Your super meter does go along way, though.
SheldonD said:
Oh I see what you mean. Yes the amp would benefit from matched current sense resistors. But in practice (tested) it is not as critical as the gain resistors. I would suspect even 5% would be just fine for those, I will have to do more research on that. To be safe we are using at least 1%.
BTW, my DMM is not anything special.
its an old Keithley beast I picked up on eBay. But it is dependable. You could probably get one for about $50Cheers!
Russ
resistor matching
when checking the resistors is the 0.1 spec a percentage of the total resistors value or percentage of 1 ohm. when i wire the tpm in bridge do + outputs from txd go to one side and + in/ and the - output go to + in on the other set. one last question, what is the package number for the lm3886 that lays flat parallel to pcb. with the small compact size of these boards i could mount the pcb flat to the sink. would make nice neat setup.
when checking the resistors is the 0.1 spec a percentage of the total resistors value or percentage of 1 ohm. when i wire the tpm in bridge do + outputs from txd go to one side and + in/ and the - output go to + in on the other set. one last question, what is the package number for the lm3886 that lays flat parallel to pcb. with the small compact size of these boards i could mount the pcb flat to the sink. would make nice neat setup.
The 0.1% is in relation to each other. The specific value isn't as important; it's making sure both paths look identical electrically.
When using the TPMs with a TXD, use only the +Output of each. The TPM driven by the TXD's + is +, and the one driven by the TXD's - out is -. Hope that makes sense (check the wiring diagram in the manual).
There is no package that I am aware of that has pins bent that way. People do this by custom-bending the pins.
When using the TPMs with a TXD, use only the +Output of each. The TPM driven by the TXD's + is +, and the one driven by the TXD's - out is -. Hope that makes sense (check the wiring diagram in the manual).
There is no package that I am aware of that has pins bent that way. People do this by custom-bending the pins.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.