I tested the FET version. It sounds great and there is no offset on start up. I don't have sophisticated testing gear though. I might be tempted to use the FET version for that reason.
I really like the Tubsumo but dealing with the offset at turn-on is extra to deal with without the 21st century board.
That's right, a power-on delay is required for the speakers for sure as soon as the tube is used at the input.
I've got the 21st century board installed and configured for the Tubsumo. What would you call a very safe delay for the tube? 25 seconds?
BTW, I was listening to the NS-OPS with CFA IPS last night and this on my cheap Bose speakers (did some calibration and needed to test first), and even with those speakers, I got immersed in the music. Great soundstage as well! That tells a lot I believe. This weekend on my Troels ATS-4 for the first time.
Ciao!
Do
BTW, I was listening to the NS-OPS with CFA IPS last night and this on my cheap Bose speakers (did some calibration and needed to test first), and even with those speakers, I got immersed in the music. Great soundstage as well! That tells a lot I believe. This weekend on my Troels ATS-4 for the first time.
Ciao!
Do
I use for tube preheating 10 seconds, i think that is enough for a little tube, and setting for a longer delay makes no difference.
What is important, is letting for the offset to settle, so you have to set the `long speakersDelay =` to 25 seconds(this is the time that the amplifier waits since you apply the power to the moment that the speakers are connected). You may experiment with shorter time if the amplifier do not enter in DC offset protection mode.
What is important, is letting for the offset to settle, so you have to set the `long speakersDelay =` to 25 seconds(this is the time that the amplifier waits since you apply the power to the moment that the speakers are connected). You may experiment with shorter time if the amplifier do not enter in DC offset protection mode.
Last edited:
Time delay setting mainly depends on the heater mode. A parallel heated tube, fed by AC from a transformer winding or by a regulated DC supply, can expected to be operative in not more than 15 seconds. A CCS fed heater needs at least half a minute or up to one minute to stabilize the operating points.
But the 21st century board also provides DC supervision at the loudspeaker output, doesn't it? So it automatically won't activate the relays as long as DC level is too high.
Best regards!
But the 21st century board also provides DC supervision at the loudspeaker output, doesn't it? So it automatically won't activate the relays as long as DC level is too high.
Best regards!
Last edited:
It's best just to measure the time it takes for offset to settle and use that value for programming. If the protection system sees DC when the speakers are ready to engage it will shut the amp down.
My set is for 25 seconds. If i set it to 20 seconds the amplifier do not enter in DC protection mode, but i hear a small click in one of the speakers when the speakers are connected---the offset is not 0 volts, but is below the threshold of DC offset detection circuit.
Please see post #2688
Measure voltages across R6,R11 compare these with yours.
This indicate how close match the double triode.
Measure voltage on LF411 pin 6 ,this indicate the servo action(the voltage that is necessary for zero offset setting).
Last edited:
Ok Thimios, i will do this tomorrow with the channel that has the biggest initial offset.
Jeff,
Wouldn't be a good idea to add some programming to wait an extra X amount of time only upon power up if DC is detected? Let's say you set your threshold to 25 seconds for filament and once the amp is ready to kick in and sees DC over the specified level, to wait an extra amount of time (something you can define up to a certain safety value) and if this gets fixed before that timeout than the speaker protection disengage itself. If on the other end the second DC threshold is expired and DC is still over safety levels, shut the amp down?
Thanks
Do
Wouldn't be a good idea to add some programming to wait an extra X amount of time only upon power up if DC is detected? Let's say you set your threshold to 25 seconds for filament and once the amp is ready to kick in and sees DC over the specified level, to wait an extra amount of time (something you can define up to a certain safety value) and if this gets fixed before that timeout than the speaker protection disengage itself. If on the other end the second DC threshold is expired and DC is still over safety levels, shut the amp down?
Thanks
Do
We could do something like this but it's not really necessary. Once the tube warms up the offset settles quickly and the time require doesn't really change. measuring the time it takes to settle out and entering that value will give you about 10 seconds of extra time while the supply soft starts and the speaker relay delay times out.
I have the measurements.
After the startup sequence and after offset reaches 0V i have:
13.5V across R6
13.5V across R11
Pin 6 voltage referenced to ground 6.7V
After the startup sequence and after offset reaches 0V i have:
13.5V across R6
13.5V across R11
Pin 6 voltage referenced to ground 6.7V
Right - depends on the algorithm (that depends on the general approach 🙂).
In our case, during the soft-start sequence, we check the offset right before engaging the speakers - if it's not within the limits, we shut down the amp, assuming something is wrong, and indicating the offset is out of range (conservative, safety-focused approach).
In my prototype of Tubsumo, I used this set of parameters:
const int TubesAreHere = 1;// 1 = Tubes, 0 = NO Tubes
long heatingDelay = 25000; // wait for tubes pre-heating (mS)
long inrushDelay = 5000; // wait for soft start (mS)
long overlapDelay = 500; // overlap before inrush goes off, if "0" - stays on (mS)
long speakersDelay = 15000;// wait before connecting speakers (mS)
Thimios's thinking is right - this most likely means the tube's triodes are not matched enough - the servo works pretty hard to keep the circuit in balance. Not critical, but better-matched triodes are preferred for the lowest distortion. The ones used by Thimios are very good - 1.14V at pin 6 is low enough.
Most of 12AU7 datasheets say the Heater warm-up time = 11 seconds. So, 25s is a safe value, I think.