DHT simulation? Haven't heard that term before. Tried searching but couldn't find an definition. Can you explain?
Brian
DHT = Distorção Harmonica Total (sorry it is in my mother laguage)
THD - is the term in english - Total Harmonic Distortion.
I corrected it in my previous message.
Ronaldo
DHT = Distorção Harmonica Total (sorry it is in my mother laguage)
THD - is the term in english - Total Harmonic Distortion.
I corrected it in my previous message.
Ronaldo
Last edited:
Somebody said board design? Im up for it if you need help 🙂 ! SMD THT anything you want.
Merola,
Not sure where this type of a snide comment comes from. If you know my work, I always test all of my boards to make sure they work before releasing them for a general group buy. Exceptions are for experienced builders volunteering to be a beta tester.
Regarding the Aurum-X amp, au contraire, I have bought the PCBs, BOM, dozen Exicon latFETs, and dual 600VA trafos already. Just waiting for a free block of time to start. Yes, it’s a big chunk of change for this amp, no doubt about it.
Last edited:
Dear XRK971,
Prove it. You are showing only bare board.
Where is the amplifier working in a test bench?
The original simulation files showed problems in square wave test 10 years ago. I am not sure the changes will make effect in practice.
I believe the original problem need more care than small simple changes as was done and it was the reason to NO model files added to project as was asked in the time of this thread.
I and others followed your thread for a long time and gave up to see results - so it is not a snide comment, but a fact.
One of my old teacher used to say: 'lack of time is the excuse of those who waste time for lack of method'
Prove it. You are showing only bare board.
Where is the amplifier working in a test bench?
The original simulation files showed problems in square wave test 10 years ago. I am not sure the changes will make effect in practice.
I believe the original problem need more care than small simple changes as was done and it was the reason to NO model files added to project as was asked in the time of this thread.
I and others followed your thread for a long time and gave up to see results - so it is not a snide comment, but a fact.
One of my old teacher used to say: 'lack of time is the excuse of those who waste time for lack of method'
Last edited:
r_merola - Your above post sure seems out of place on a DIY forum with people having fun in their hobby... Please do us all a favor and relax a little and enjoy your hobby. 😀
Dear 6L6
All I wrote about XRK971 was facts that can be followed in this forum. It is just an advice to avoid troubles instead of fun in this hobby.
I also gave Ricardo information to improve his work.
I will relax...
All I wrote about XRK971 was facts that can be followed in this forum. It is just an advice to avoid troubles instead of fun in this hobby.
I also gave Ricardo information to improve his work.
I will relax...
Last edited:
Keep calm and carry on.
I also did some square wave sims and the results are very good.
I normally use 50khz squares and probe around ips and vas looking for weird current shapes.
I do use squares for fine tunning after building the amps 🙂
I also did some square wave sims and the results are very good.
I normally use 50khz squares and probe around ips and vas looking for weird current shapes.
I do use squares for fine tunning after building the amps 🙂
Yes, 8db gain margin is too short... will work on that... thank you for your analysys
No
You may use matched sets (which you can get from profusion and I do) however, even this is not necessary. Some people match the gate resistances through measurement and fine tuning tho, another unnecessary thing beyond the fact that at times, the P and N devices have different gate capacitances which can be compensated for with different gate resistances. You will notice that the ECX10N20 and ECX10P20 have a 500pF gate capacitance and the ECW20P20 dual-die has 1850pF, while the ECW20N20 has 900pF. I use the dual units and have to double the gate resistance of the N device. Other makes like the Hitachi also have different gate capacitances. Some designers just add capacitance to the lesser device.
And as a further... You may use source resistors to set your bias, and then check your voltage drop across the bias trimmer and use this as your new method of setting the bias without source resistors. Once you get a ballpark resistor value, you can omit the trimmer and solder in a fixed resistor. If there is a group of people doing these, there can be some discussion as to the best resistor value and then blindly place that, trusting the unit, and just checking for DC at the output and perhaps the drop across the bias resistor for that ballpark value. In general, there are many designs that it nearly depends not at all on the bias setting. My units, along with Rod Elliot will profess that his had nearly no effect and has outstanding results. I would say find your ballpark target and then just go with a fixed resistor.
S
S
From the manufacturer..
"Freedom from secondary breakdown and thermal runaway make them extremely reliable and remove the need for protection circuitry.
When I get my 3rd proto finished, I plan to test it with a 2ohm sub for as long as it will play full power, until the devices throttle themselves. I've achieved this on a different unit with a higher rail V and perhaps smaller heatsink. I have no protection on my current proto and have run 2 4ohm full range on a channel for a successful period. My goal is to have stable 2ohm for at least a short term playback.
S
Thanks synonymous
Could you also set the bias without source resistors by shorting the input and checking the current draw from each rail? Seems you could initially set the bias to zero and read the current draw from the rest of the circuit for each rail, then trim it out to the desired bias. If there are multiple pairs, this would assume they are sharing the bias current equally. I assume you could get a ballpark validation of current sharing (or lack thereof) by using an IR thermometer on the case of each device.
Would this approach work?
Could you also set the bias without source resistors by shorting the input and checking the current draw from each rail? Seems you could initially set the bias to zero and read the current draw from the rest of the circuit for each rail, then trim it out to the desired bias. If there are multiple pairs, this would assume they are sharing the bias current equally. I assume you could get a ballpark validation of current sharing (or lack thereof) by using an IR thermometer on the case of each device.
Would this approach work?
I wouldn't get too creative in the method if I was anybody.
You can use nearly any resistor type and physical size to temporarily fill in as source resistors for setting it. You don't even need a place on the circuit board for them if you dont insert the source lead. Just temporarily use another resistor. See attachment.
S
You can use nearly any resistor type and physical size to temporarily fill in as source resistors for setting it. You don't even need a place on the circuit board for them if you dont insert the source lead. Just temporarily use another resistor. See attachment.
S
The voltage I use across Rtrim is between 0.85V and 0.89V
That should ballpark you and possibly even be your solution.
Erring here would be to use 0.85V as your low, and using a fixed resistor, it would be okay to get a slightly higher drop.
I do recommend using temporary source resistors to double check your work and secure your mind.
After all this, you can, through multiple test units, come up with a resistance that may work for all of them and everyone else.
S
That should ballpark you and possibly even be your solution.
Erring here would be to use 0.85V as your low, and using a fixed resistor, it would be okay to get a slightly higher drop.
I do recommend using temporary source resistors to double check your work and secure your mind.
After all this, you can, through multiple test units, come up with a resistance that may work for all of them and everyone else.
S
I won't call this creative. Nelson Pass mentions this method in the F6 Burning Amp talk paper. When discussing the Power Supply CRC filter, he states:
Assuming you use a CRC filter in your power supply, you can measure the voltage drop across the filter resistor. If its a 0.1Ω resistor, multiply the mV value by 10 to get your current in mA through the rail. Back out the IPS, VAS and Driver currents to arrive at the OPS bias current.
Assuming you use a CRC filter in your power supply, you can measure the voltage drop across the filter resistor. If its a 0.1Ω resistor, multiply the mV value by 10 to get your current in mA through the rail. Back out the IPS, VAS and Driver currents to arrive at the OPS bias current.
If you want to try that technique, feel free to do so. I would still double check the results using source resistors initially.
I am only saying this for your own sake, to know for yourself how accurate each method corresponds to another.
Once you know some data, you can do as i do (since I use neither source resistors or CRC), and just measure for 0.85V or thereabouts, across Rtrim.
I just ran a test and checked my Rtrim for both R+L channels and found one to be trimmed to 86 Ohm and another 93 Ohm. So as you can see, a 90 Ohm fixed resistor would do the trick just fine for both channels.
S
I am only saying this for your own sake, to know for yourself how accurate each method corresponds to another.
Once you know some data, you can do as i do (since I use neither source resistors or CRC), and just measure for 0.85V or thereabouts, across Rtrim.
I just ran a test and checked my Rtrim for both R+L channels and found one to be trimmed to 86 Ohm and another 93 Ohm. So as you can see, a 90 Ohm fixed resistor would do the trick just fine for both channels.
S
Last edited: