I no longer have DC offset from the time I put the jumper on the board. The square wave is the output of the amp; 0.62 V point to point and according to my multimeter is says 0.1 mV DC.
Yesterday I also made it drive some current. The pack of "rail resistors" were magically transformed into one 30 Watt 12 Ohm resistor. I have proven the device can do more than an OPA2132 can, drive current!
I will take this thing to a different place where they have nice psu's and better signal generators. Any measurements I should do in order to get a nice view of how the amp is performing?
I no longer have DC offset from the time I put the jumper on the board. The square wave is the output of the amp; 0.62 V point to point and according to my multimeter is says 0.1 mV DC.
Yesterday I also made it drive some current. The pack of "rail resistors" were magically transformed into one 30 Watt 12 Ohm resistor. I have proven the device can do more than an OPA2132 can, drive current!
I will take this thing to a different place where they have nice psu's and better signal generators. Any measurements I should do in order to get a nice view of how the amp is performing?
Can you access a distortion analyzer and/or FFT (spectrum) analyzer? Have you adjust the bias level?
I suggest you download the book "Audio Measurement Handbook" from AP AP High Performance Audio Analyzer & Audio Test Instruments : Downloads in order to know what do measure.
Make sure the signal generator output is DC free. Otherwise, you need to insert an input DC block cap.
Thermal tracking exercise
Here I continue my last experiment using two diodes attached on the main heatsink for thermal tracking (one shown in the picture). The two diodes' forward voltage, a 500R trim pot and ThermalTrak diodes set the bias level.
The initial bias level was adjusted for 23 mA passing through the output devices (5 mV across 0.22R emitter resistor). The heatsink temp is 30 C when the amp is idle (room temp = 28 C). Input signal was then applied for output 36 W into a 4 Ohm resistive load. The input was removed when the heatsink temp reached 50 C. After 10 sec, the Re drop was 6.4 mV (29 mA), heatsink temp was 48 C. When the heatsink was 42 C, the Re drop was 6.0 mV (27.2 mA). When heatsink was 37 C, Re drop was 5.6 mV (25.5 mA). The tracking performance is still not "perfect", but much better than only using TT diodes. More effort needed.
Here I continue my last experiment using two diodes attached on the main heatsink for thermal tracking (one shown in the picture). The two diodes' forward voltage, a 500R trim pot and ThermalTrak diodes set the bias level.
The initial bias level was adjusted for 23 mA passing through the output devices (5 mV across 0.22R emitter resistor). The heatsink temp is 30 C when the amp is idle (room temp = 28 C). Input signal was then applied for output 36 W into a 4 Ohm resistive load. The input was removed when the heatsink temp reached 50 C. After 10 sec, the Re drop was 6.4 mV (29 mA), heatsink temp was 48 C. When the heatsink was 42 C, the Re drop was 6.0 mV (27.2 mA). When heatsink was 37 C, Re drop was 5.6 mV (25.5 mA). The tracking performance is still not "perfect", but much better than only using TT diodes. More effort needed.
Attachments
Panson, how would you describe the amplifiers sound? I am playing my first tunes on a single channel (old test loudspeakers) but the sound seems a bit "shallow/flat". it doesn't lack bass but it lacks "body". I am not sure if that is due to the testspeaker or the amplifier.
Second. while playing the mje15030's get around 35 degrees Celcius, the LME49811 gets up to 45 degrees Celcius and the NJL's temperature won't rise above 25 degrees Celcius. Is that similar to your tests? They are mounted on a heatsink but I would have expected them to get hotter.
Second. while playing the mje15030's get around 35 degrees Celcius, the LME49811 gets up to 45 degrees Celcius and the NJL's temperature won't rise above 25 degrees Celcius. Is that similar to your tests? They are mounted on a heatsink but I would have expected them to get hotter.
What is the bias level do you currently use? If the output device bias level is 22 mA as used in my test, they will not get hot for low level output. Do you have heatsink for LME and driver? The temp of them seems normal. You can try to test the amp with a power resistor load. Let it run for 1/3 rated power for, say, 30 min. The output devices should then be very warm.
Please try stereo setup with a nice speaker for sound evaluation. You signal source should be high quality too.
Bias is 20 mV over 2*0.18Ohm so +/- 50mA on the 1st channel
Bias is 40 mV over 2*0.18Ohm so +/- 100mA on the 2nd channel
I can't hear a difference between the two channels. I can easily turn them lower/higher but what is recommended and most of all, why is it recommended? Googling doesn't make thing
If I wouldn't care about power consumption and have large heatsinks. Should I stick with this 50-100mA range?
Bias is 40 mV over 2*0.18Ohm so +/- 100mA on the 2nd channel
I can't hear a difference between the two channels. I can easily turn them lower/higher but what is recommended and most of all, why is it recommended? Googling doesn't make thing
If I wouldn't care about power consumption and have large heatsinks. Should I stick with this 50-100mA range?
If you can not hear any difference in your system, go for lower bias level. If you have a FFT analyzer or spectrum analyzer, you can see the harmonic structure changing with bias level.
Now, almost 3 weeks later on, I put together my loudspeaker protection and connected my normal speakers. Here is the new "problem"/"challenge".
I have a slight hiss when I connect my cd-player's headphone output (with volume regulation). I have the same hiss when I connect my iPod. Both are noticeable when volume = 0 or a song is paused. I need to put my ears close to the tweeter, but we all want amps to be 100% silent@pause! The hiss disappears when I short the amp's input.
What is wrong and more important, how to get rid of the hiss? Is it an impedance issue which should be solved by the use of a 'normal stereo log potmeter' or should I buffer it with a 1:1 opa2132?
I am nearly there before judging the sound quality, but as long as the amp isn't functioning as I would like, I won't test it just yet.
I have a slight hiss when I connect my cd-player's headphone output (with volume regulation). I have the same hiss when I connect my iPod. Both are noticeable when volume = 0 or a song is paused. I need to put my ears close to the tweeter, but we all want amps to be 100% silent@pause! The hiss disappears when I short the amp's input.
What is wrong and more important, how to get rid of the hiss? Is it an impedance issue which should be solved by the use of a 'normal stereo log potmeter' or should I buffer it with a 1:1 opa2132?
I am nearly there before judging the sound quality, but as long as the amp isn't functioning as I would like, I won't test it just yet.
Shorting the input terminals is a quick way to determine that the hiss does not originate in the amplifier. Try putting a volume pot inbetween the amp and your input devices. Run those devices at maximum output, and control the main volume with your new pot.
The reason you are getting hiss is the poor signal-to-noise ratio that results from controlling the volume in your input devices.
The reason you are getting hiss is the poor signal-to-noise ratio that results from controlling the volume in your input devices.
Last edited:
That helped much more than I expected it would. I am playing my first tunes in stereo on my normal loudspeakers! First impressions, way to early to tell but I don't hear any abnormalities. So far so good. Btw, the heatsing for the output devices get way hotter then I previously thought, +/- 50 degrees Celcius with 22 gedrees Celcius ambient.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Mini update;
The heatsinks kept getting hotter to my surprise. When the thermometer hit >60 degrees Celcius and I couldn't hold the heatsinks for more than a second, I decided it was time to put the "fins" vertically so they could dissipate more heat. As I was quite interested in the bias at these high temperatures I put the cd player on pause and measured the voltage over the 2*0.18 Ohms I used. Prepare yourself, the initial value was +/- 30 mV at ambient. At the high temperature it went up to a whopping 70mV, 200mA bias!
After the first 2 cd's I might conclude it sounds better than my Denon PMA500V in terms of "stereo image" and "bass control". However, the measurements above can't be normal. With the fins vertically and the bias back to the <100mA range the heatsinks get 40-45 degrees Celcius.
I remember Panson telling me he used a copper bar and the aluminium of his casing for cooling...
The heatsinks kept getting hotter to my surprise. When the thermometer hit >60 degrees Celcius and I couldn't hold the heatsinks for more than a second, I decided it was time to put the "fins" vertically so they could dissipate more heat. As I was quite interested in the bias at these high temperatures I put the cd player on pause and measured the voltage over the 2*0.18 Ohms I used. Prepare yourself, the initial value was +/- 30 mV at ambient. At the high temperature it went up to a whopping 70mV, 200mA bias!
After the first 2 cd's I might conclude it sounds better than my Denon PMA500V in terms of "stereo image" and "bass control". However, the measurements above can't be normal. With the fins vertically and the bias back to the <100mA range the heatsinks get 40-45 degrees Celcius.
I remember Panson telling me he used a copper bar and the aluminium of his casing for cooling...
Last edited:
Dear SuperR,
I think 2 things. I think the combination of relative low ohmic output resistors and the relative small heat-sink in combination with your high bias levels result in thermal runaway. I think your bias levels are to high and suggest you lower them between 20 and 25mV over each Re.
higher output resistors like 0.22Ohm or 0.33Ohm gives a bit more thermal stability.
Also if you adjust the bias, let the amplifier warm up first, then adjust, wait again and adjust.
Keep in mind also, if you measure bias just after a period of (loud) listening levels, the amplifier and bias need time to go back to idle. Normal this takes about a minute of 1 or 2.
Also a question to Panson. Isn't it that only two termaltrak's have not enough diodes for accurate thermal tracking?
with kind regards,
Bas
Last edited:
Try not to take the signal from the headphone output of your CD player. Use the line output. Headphone amp of the CD player and your iPod are much noisy source for amplification.
Yes, it is. This is why I use two extra diodes to get better thermal tracking as mentioned in post #145 of this thread. The result is not perfect, but much better.
Hey Panson...
Your design caught my eye
What about using more output devices instead of adding more diodes ??
Sorta like this ...
Your design caught my eye
What about using more output devices instead of adding more diodes ??
Sorta like this ...
An externally hosted image should be here but it was not working when we last tested it.
Hey Panson...
Your design caught my eye
What about using more output devices instead of adding more diodes ??
Sorta like this ...
An externally hosted image should be here but it was not working when we last tested it.
Hi Adrculda,
This is actually what I am doing here http://www.diyaudio.com/forums/chip-amps/143163-comparing-lme49810-49830-49811-a.html I have used two and four pairs.
How is your design doing?
Cheers,
Panson
Last edited:
Still on Paper, but parts are on their way
And its actually a copy of yours with 2 more sets of output devices.
I got 2 x 750VA Plitron Toroids and about 2 Farads worth of caps laying around here waiting for my behind to get started on the power supply's.
I'm ALMOST tempted to see how hard would it be to make a multi rail supply sorta like a class G or H...
