Inverted OTL with 6SN7/6AS7
#46 Re: Inverted OTL with 6SN7/6AS7
Well, actually what I would so is decide what output from the amp I would think is a good value. 1v into 8 ohms is 125mW, that may be a good value for the OTL. I would inject 1v RMS into the scope, look at the FFT and look a the difference between the 1kHz spike and the surrounding noise floor. So say spike is at -15dB, surrounding floor is -85dB, so the measuring noise floor is at 70dB below 1v, so that means that -85dB on the scope is 316uV RMS.
Then connect to the amp, adjust the 1kHz signal from the amp so it is at the same -15dB (or 1v RMS however its easier to measure), then look at the surrounding noise floor, and if for example its at -65dB, we can then say that the amp floor is at 3mV
Its not a measure of the noise, but that would need a bandwidth filter and a true RMS meter, to allow you to give a figure in V per Root Hz. Though I think that will only be true if the noise is actually noise in a Gaussian sense.
What I find is actually useful instead of just creating numbers is to look at the level of that background noise, and the distribution against frequency, and of course any power supply related spikes. But its still important to have a idea of the voltage levels that the display is showing you.
Then connect to the amp, adjust the 1kHz signal from the amp so it is at the same -15dB (or 1v RMS however its easier to measure), then look at the surrounding noise floor, and if for example its at -65dB, we can then say that the amp floor is at 3mV
Its not a measure of the noise, but that would need a bandwidth filter and a true RMS meter, to allow you to give a figure in V per Root Hz. Though I think that will only be true if the noise is actually noise in a Gaussian sense.
What I find is actually useful instead of just creating numbers is to look at the level of that background noise, and the distribution against frequency, and of course any power supply related spikes. But its still important to have a idea of the voltage levels that the display is showing you.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
#47 Re: Inverted OTL with 6SN7/6AS7
How I see it now after this little discussion:
Andrew's original way would show the different noise floors without a reference. It's not possible to say anything about the absolute numbers as it only shows a ratio. But it would tell us how much this ratio is.
Nick's way would show us in addition also absolute numbers as he has created a reference with the -15dB signal strength.
SNR measurement would show us only the ratio between noise floor and signal but without information about the noise floor of the scope.
I think absolute numbers would only be interesting if we measure things like a DAC for example as in this case the absolute numbers show something about it's resolution/quality. In the case of an amp absolute numbers are not important as long as we only want to find out about possible problems.
I measured with a 1kHz signal(-18dB) and the noise floor to signal is 50dB! Or I could simply say the noise floor is 35% higher than without using a signal.That's already something. Good to know.
Andrew's original way would show the different noise floors without a reference. It's not possible to say anything about the absolute numbers as it only shows a ratio. But it would tell us how much this ratio is.
Nick's way would show us in addition also absolute numbers as he has created a reference with the -15dB signal strength.
SNR measurement would show us only the ratio between noise floor and signal but without information about the noise floor of the scope.
I think absolute numbers would only be interesting if we measure things like a DAC for example as in this case the absolute numbers show something about it's resolution/quality. In the case of an amp absolute numbers are not important as long as we only want to find out about possible problems.
I measured with a 1kHz signal(-18dB) and the noise floor to signal is 50dB! Or I could simply say the noise floor is 35% higher than without using a signal.That's already something. Good to know.
#48 Re: Inverted OTL with 6SN7/6AS7
On the other hand I would suggest that relative numbers would only be of interest if you were comparing two similar things or the same amp before and after a component change. Being able to say your amp has 15dB more noise than your picoscope tells you nothing of real interest about the amp (or the picoscope for that matter).I think absolute numbers would only be interesting if we measure things like a DAC
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
#49 Re: Inverted OTL with 6SN7/6AS7
Time to move on to the next step.
I am almost there that I can put both channels in the amp housing I was building (wood) and I was thinking about how to solve practically the remaining issues with this amp.
1. How to deal with 40V DC at the output after turning the amp on until everything has evened out (about 15s for the voltage to be down at DC offset, about 10min for DC offset to be down at the lowest adjusted voltage)
2. How to deal with the DC offset fluctuations between 20mV and 400mV depending on the settings and temp of the tubes. This is different than with a PP amp. Controlling the grids for stable DC offset is somehow easier than controlling the anode. Never the same twice with the InvOTL but it always comes back pretty close to the adjusted DC offset with the PP OTL when the ideal setting has been found. No matter how I adjust the 250k pot DC offset will always be off quite a bit when I turn it on the next time.
The easiest solution for #1 would be to use an output cap. All problems solved. But the thrill with this amp is the sound and it suffers a subtle but noticeable loss with an output cap (A/B comparison).
Don't know what S.Bench had in mind with his original InvOTL. It operates at lower psu voltages at the output with the 5687 tubes but that wouldn't make much difference as I can see the membrane of an PM6A already moving a few mm at a DC offset of 100mV. Maybe less sensitive speakers with a x-over would make it less critical? But not with 40V and the 6AS7 version.
So this is what I will use for the turning on/turning off sequence. It's easy to build and 100% reliable. The values of the resistors for setting the time need to be still fine tuned for the final settings.
In order to solve issue #2 I followed Nick's advice and used an opamp in combination with an LDR and got some interesting results. But I need to test the circuit first with the actual amp. It could work and help to control the DC offset.
I am almost there that I can put both channels in the amp housing I was building (wood) and I was thinking about how to solve practically the remaining issues with this amp.
1. How to deal with 40V DC at the output after turning the amp on until everything has evened out (about 15s for the voltage to be down at DC offset, about 10min for DC offset to be down at the lowest adjusted voltage)
2. How to deal with the DC offset fluctuations between 20mV and 400mV depending on the settings and temp of the tubes. This is different than with a PP amp. Controlling the grids for stable DC offset is somehow easier than controlling the anode. Never the same twice with the InvOTL but it always comes back pretty close to the adjusted DC offset with the PP OTL when the ideal setting has been found. No matter how I adjust the 250k pot DC offset will always be off quite a bit when I turn it on the next time.
The easiest solution for #1 would be to use an output cap. All problems solved. But the thrill with this amp is the sound and it suffers a subtle but noticeable loss with an output cap (A/B comparison).
Don't know what S.Bench had in mind with his original InvOTL. It operates at lower psu voltages at the output with the 5687 tubes but that wouldn't make much difference as I can see the membrane of an PM6A already moving a few mm at a DC offset of 100mV. Maybe less sensitive speakers with a x-over would make it less critical? But not with 40V and the 6AS7 version.
So this is what I will use for the turning on/turning off sequence. It's easy to build and 100% reliable. The values of the resistors for setting the time need to be still fine tuned for the final settings.
In order to solve issue #2 I followed Nick's advice and used an opamp in combination with an LDR and got some interesting results. But I need to test the circuit first with the actual amp. It could work and help to control the DC offset.
#50 Re: Inverted OTL with 6SN7/6AS7
I forgot to attach this shot of the amp modules that I will put into the housing as soon as it's done.
Like this I don't need to drill too many holes that can be seen from outside. This is kind of important in this case as I try to make the box look like intarsien wood finish as I feel that this amp deserves a special outer finish. I hope it works out.
Like this I don't need to drill too many holes that can be seen from outside. This is kind of important in this case as I try to make the box look like intarsien wood finish as I feel that this amp deserves a special outer finish. I hope it works out.
- Paul Barker
- Social Sevices have been notified
- Posts: 8988
- Joined: Mon May 21, 2007 9:42 pm
#51 Re: Inverted OTL with 6SN7/6AS7
Looks organised.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
#52 Re: Inverted OTL with 6SN7/6AS7
Nice, though I have got lazy now, I would have just used as ATTiny24.So this is what I will use for the turning on/turning off sequence. It's easy to build and 100% reliable.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
#53 Re: Inverted OTL with 6SN7/6AS7
quite agree, although just look at the elegance in the old school approach, and the satisfaction...I'm on the fence with this one.
There's nowhere you can be that isn't where you're meant to be
#54 Re: Inverted OTL with 6SN7/6AS7
Though its behavior with a short term power out may be indeterminate unlike a controller with a brown out detector and known reset state.100% reliable
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
#55 Re: Inverted OTL with 6SN7/6AS7
I think I'm off the fence now..that's a very good point about the power out. I haven't run throungh the complete nand gate logic path but I think in principal it's the same scenario with the class d mute problem..i.e when the power comes on the mute must be on to avoid thumps but when the power goes off the mute must also be on to avoid the off thump....
I'm not sure if it's laziness or practicality but in that scenario I'm with the avr...it's all about who gets to the switch first and if there is some residual in the power supply to keep going while you run for the switch...
I'm thinking it's the same problem maintaining equilibrium in a standard push pull set up when both sides dont' bias up or down at the same rate.
I'll stop rambling now cos I only think I undertsand what's going on here
I'm not sure if it's laziness or practicality but in that scenario I'm with the avr...it's all about who gets to the switch first and if there is some residual in the power supply to keep going while you run for the switch...
I'm thinking it's the same problem maintaining equilibrium in a standard push pull set up when both sides dont' bias up or down at the same rate.
I'll stop rambling now cos I only think I undertsand what's going on here
There's nowhere you can be that isn't where you're meant to be
#56 Re: Inverted OTL with 6SN7/6AS7
The secondaries which control the timer part are buffered with 30kuF which keeps the SSRs functional at power outage until they have shut down in the preset sequence. Muting the output comes always first so whatever happens (short interruption or real power outage)there is never danger for the speakers. The only thing to consider is that the relais which switch the power should draw only little current and/or the buffer capacity needs to be increased accordingly . Flickering mains power will not trigger the full sequence and the speakers will stay disconnected under all circumstances.Though its behavior with a short term power out may be indeterminate unlike a controller with a brown out detector and known reset state.
If one doesn't use a PCB it would be hard to use the SMD ATtiny24. Also programming the eprom might not be easy if one isn't used to do this kind of stuff.
#57 Re: Inverted OTL with 6SN7/6AS7
But why would they shut-down in the case of a power outage?until they have shut down in the preset sequence.
I use the DIP16 version, same as the gates you are using.If one doesn't use a PCB it would be hard to use the SMD ATtiny24.
Yep, though its very simple. I wasn't suggesting that you should change, just saying why I would be lazy.Also programming the eprom might not be easy if one isn't used to do this kind of stuff.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
#58 Re: Inverted OTL with 6SN7/6AS7
Because Rel 5 turns off and the output of IA 1/2/3 goes to 0. The caps will lose their charging according to the time settings with resistor/no resistor and the SSRs will turn off.But why would they shut-down in the case of a power outage?
They are indeed still available.I use the DIP16 version, same as the gates you are using
#59 Re: Inverted OTL with 6SN7/6AS7
Ahh I see, clever.Because Rel 5 turns off
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
#60 Re: Inverted OTL with 6SN7/6AS7
While I was getting ready for putting everything into the amp housing I realized that there is a mistake in the schematics of the output power supply.
The two 20ohm/20W resistors between the two caps have to be in series (40ohm) and NOT parallel as the schematics show. It must have happened when I changed from resistor to choke loaded grids.
The two 20ohm/20W resistors between the two caps have to be in series (40ohm) and NOT parallel as the schematics show. It must have happened when I changed from resistor to choke loaded grids.