How about this: http://www.st.com/stonline/products/literature/ds/9921/stth506d.pdf
and this: http://ixdev.ixys.com/DataSheet/L373.pdf
and this: http://ixdev.ixys.com/DataSheet/L373.pdf
Last edited:
Just for info.
Back EMF is going to be 3 or 4 times the value of the supply on the diodes!
The PIV ( Peak inverse voltage must be greater than this value)
If you cannot get a higher value of PIV then try putting 2 diodes in series to replace each diode!
You can try MOV, however MOV degrade with time and operations and are only meant to "take out" ramdom spikes! The value and type of MOV must be able to withstand the current when it breaks down.
It is quite common to see 1000-2000 Volts back EMF at power off.
The type of diodes you use will effect the noise floor of the amp!
You can try soft recovery as well as HEX. All you are doing is removing the switching noise of the diodes. ( The HT supply is part of the signal path thats why you hear a difference!)
The back EMF does not do your tubes much good either! You may exceed the max working voltage of the tube!
I have also seen MOV's used on the Primary as well as secondary to help remove this problem.
Regards
M. Gregg
Back EMF is going to be 3 or 4 times the value of the supply on the diodes!
The PIV ( Peak inverse voltage must be greater than this value)
If you cannot get a higher value of PIV then try putting 2 diodes in series to replace each diode!
You can try MOV, however MOV degrade with time and operations and are only meant to "take out" ramdom spikes! The value and type of MOV must be able to withstand the current when it breaks down.
It is quite common to see 1000-2000 Volts back EMF at power off.
The type of diodes you use will effect the noise floor of the amp!
You can try soft recovery as well as HEX. All you are doing is removing the switching noise of the diodes. ( The HT supply is part of the signal path thats why you hear a difference!)
The back EMF does not do your tubes much good either! You may exceed the max working voltage of the tube!
I have also seen MOV's used on the Primary as well as secondary to help remove this problem.
Regards
M. Gregg
Regarding back EMF,
Any coil (including chokes) that have DC across them will produce a back EMF (voltage higher than the suppy to them in the opposite direction to applied voltage) so if you have + volts on the start of the winding on power on, you will have - volts (1000V EMF) on that side at power off.
So any diodes used to remove back EMF will resist the supply and have no effect until power off.
EMF is produced by the magnetic field cutting through the windings of the Tx as it Collapses.
Just for fun
Regards
M. Gregg
Any coil (including chokes) that have DC across them will produce a back EMF (voltage higher than the suppy to them in the opposite direction to applied voltage) so if you have + volts on the start of the winding on power on, you will have - volts (1000V EMF) on that side at power off.
So any diodes used to remove back EMF will resist the supply and have no effect until power off.
EMF is produced by the magnetic field cutting through the windings of the Tx as it Collapses.
Just for fun
Regards
M. Gregg
Just for interest,
"Conventional and non conventional current flow".
Note that electrons move from - to + in the valve (Correct), however the diode direction of flow symbol shows + into the diode - out. So you have to recognise this when looking at circuits showing both types of component!
Regards
M. Gregg
"Conventional and non conventional current flow".
Note that electrons move from - to + in the valve (Correct), however the diode direction of flow symbol shows + into the diode - out. So you have to recognise this when looking at circuits showing both types of component!
Regards
M. Gregg
There are transient ringing effects during switch off that are much worse with plain jane 1N4007 compared to UF4007. These high frequency transients occur at 2X the line frequency. Thus you get low frequency (2X line) and high frequency noise in your PS.
These get coupled into the gain stages and add to floor noise which impacts imaging.
I first experienced this with a PP 6P1P amp. For some reason it only effected one channel (layout related?). When I had GNFB hooked up I could not hear it, but when I was experimenting with reducing the GNFB and removed it completely I could really hear the noise. Replacing the 1N4007s with UF4007s got rid of the noise and improved the overall quality of the sound imaging.
These get coupled into the gain stages and add to floor noise which impacts imaging.
I first experienced this with a PP 6P1P amp. For some reason it only effected one channel (layout related?). When I had GNFB hooked up I could not hear it, but when I was experimenting with reducing the GNFB and removed it completely I could really hear the noise. Replacing the 1N4007s with UF4007s got rid of the noise and improved the overall quality of the sound imaging.
For those that like to get more involved in diode switching here is a link
Diode Turn-On Time
All just for fun
Regards
M. Gregg
Diode Turn-On Time
All just for fun
Regards
M. Gregg
Hi Gregg,
Thanks a lot for the link.
It seems the 'problems' happen in both turn on and turn off. However for rectification (of mains frequency and sine wave), the on-off transition wouldn't be so abrupt (as the test pules). So I'm wondering how they bahave in such situation.
In the book of Morgan Jones, there're some diagrams showing the waveforms of rectifying. As I see them, the glitches seem in the turn-on stage - or shall I see it as the result of slow turn-off of the other side?
In those tests, it looks best on shottky. However they are not enough PIV for tube circuit usage
Thanks a lot for the link.
It seems the 'problems' happen in both turn on and turn off. However for rectification (of mains frequency and sine wave), the on-off transition wouldn't be so abrupt (as the test pules). So I'm wondering how they bahave in such situation.
In the book of Morgan Jones, there're some diagrams showing the waveforms of rectifying. As I see them, the glitches seem in the turn-on stage - or shall I see it as the result of slow turn-off of the other side?
In those tests, it looks best on shottky. However they are not enough PIV for tube circuit usage
I must confess Ive never experienced the problem, maybe previously as an RF engineer I've always worked with proven decoupling procedures and care with layouts. The proliferation of bandit designed switchmode appliances at home is more troublesome, the noise these produce is high. I notice in my CD player it has conventional IN4003's fitted........and yet the A weighted noise floor is way below -100dB from 1V o/p. That's nothing. I can only assume that if the noise effects signal quality, then the ESR of the circuit decoupling caps is very poor.
richy
I believe you can get extra noise if the ESR is too low, as this creates a sharper charging pulse which then gets induced into other circuits. The solution is either to use cheaper capacitors or to minimise the loop enclosed by the charging pulses e.g. used twisted pair wiring. Twisted pair is messy to do on a PCB!
Here's a thought: is it possible that conventional diodes can work OK for a low voltage low current supply such as a CD player, but you need better diodes for high current (e.g. DC heaters) or high voltage (e.g. anode supplies)? My thinking is that any charge storage problem will be worse with high current circuits near signal inputs. High voltage circuits could be worse because the inverse voltage sweeps negative much more quickly so the stored charge is swept away quickly then the current abruptly stops; a low voltage supply gives the charge time to dissipate more slowly.
Here's a thought: is it possible that conventional diodes can work OK for a low voltage low current supply such as a CD player, but you need better diodes for high current (e.g. DC heaters) or high voltage (e.g. anode supplies)? My thinking is that any charge storage problem will be worse with high current circuits near signal inputs. High voltage circuits could be worse because the inverse voltage sweeps negative much more quickly so the stored charge is swept away quickly then the current abruptly stops; a low voltage supply gives the charge time to dissipate more slowly.
Hi Gregg,
Thanks a lot for the link.
It seems the 'problems' happen in both turn on and turn off. However for rectification (of mains frequency and sine wave), the on-off transition wouldn't be so abrupt (as the test pules). So I'm wondering how they bahave in such situation.
In the book of Morgan Jones, there're some diagrams showing the waveforms of rectifying. As I see them, the glitches seem in the turn-on stage - or shall I see it as the result of slow turn-off of the other side?
In those tests, it looks best on shottky. However they are not enough PIV for tube circuit usage
Regards the switch off point. The capacitors charge on the leading slope of the sine wave, do they remain conducting on the trailing edge? If they did your caps would discharge. So the diodes are under pressure from the stored charge at switch off! So it is not just the gentle slope of the AC that the diodes are following!
Remember that the tests only show limited types of diode (Hex and soft recovery have their own signature sound!)
Also just for interest the AC you are measuring is RMS. and the diodes will see the peak value of the supply + transients.
This is also true when you use MOV suppression. The MOV sees peak not RMS voltage! So if you measure RMS and fit a MOV with trigger point just above you will find it creates an interesting fire work display as it tries to dissipate the power of the AC signal!
Just for fun
Regards
M. Gregg
Interesting: I've always asssumed that the HV secondary side of the mains transformer i.e 350-0-350, 60VA exhibits quite an appreciable impedance/resistance around 300 ohms each half, combined with high C internal effectively squashes the dv/dt as to dampen out rectifier resonance effects. Maybe the reason I haven't come across this is because I work with HP stuff, using high VA transformers which have a fraction of secondary resistance 1.5 ohms ! and work into PFC circuits which are noisy enough.
I shall look into this a bit more. MAny thanks for the info
I shall look into this a bit more. MAny thanks for the info
Just for info,
I have seen diodes that test perfect appear to leak the charge from an inverter cap so it will not charge to any usefull voltage. Change it to UF type and it works, change it back to a new previously used type and watch it leak away again.
What a pain when you have changed the power transistor for the 3rd time to find out it is the type of diode used!
Just for fun!
Regards
M. Gregg
I have seen diodes that test perfect appear to leak the charge from an inverter cap so it will not charge to any usefull voltage. Change it to UF type and it works, change it back to a new previously used type and watch it leak away again.
What a pain when you have changed the power transistor for the 3rd time to find out it is the type of diode used!
Just for fun!
Regards
M. Gregg
Last edited:
I just modded my new amp (2A3 SE) with SS rectifier (UF4007), followed by a TV damper diode.
It was a hybrid bridge before - regular twin diode tube on the (+) and 2 SS diodes on the (-). I tried 5R4, 5AR4, and GZ37 for the tube side. 5R4 has huge voltage drop and start-up surge, so I dropped it after a short while. GZ37 has been used for the longest among the bunch. (Ah, not really 'long', only 1 or 2 weeks?)
And then I modded it wih full SS bridge by 4 UF4007 with snubber (100 nF +12 Ohm) on the secondary of power trafo (350 Vac with 20 Ohm DCR), followed by a resistor and a single TV damper diode - 6AU4.
This combo has less voltage drop then the above, so I have to re-adjust the RC component in the first stage of filtering to get a proper voltage. (Well, actually I didn't get it settled at the voltage precisely as before and still a little bit higher. Short of proper components on hand and time... )
Anyway it works OK this way. I really like the slow turn-on of the IH diode tube, and the damper diode seems even slower.
It was tested on table for about half an hour last night and started playing in the morning now for about an hour. I hate to say that, but it does sound different from the previous setup. The sound of full SS bridge + damper diode is a little bit 'stiffer' and slightly emphasized in higher mid. The previous half SS/half tube hybrid bridge sounds more laidback by comparison.
I thought the additional series resistor after the SS rectifier and RCLCLC filter shuold make all AC "issues" disappear, but it seems not entirely the case. The operation point of the amp is shifted a little, but only a few volts and several mA, should be no big deal.
Is it the snubber not properly set?
It was a hybrid bridge before - regular twin diode tube on the (+) and 2 SS diodes on the (-). I tried 5R4, 5AR4, and GZ37 for the tube side. 5R4 has huge voltage drop and start-up surge, so I dropped it after a short while. GZ37 has been used for the longest among the bunch. (Ah, not really 'long', only 1 or 2 weeks?)
And then I modded it wih full SS bridge by 4 UF4007 with snubber (100 nF +12 Ohm) on the secondary of power trafo (350 Vac with 20 Ohm DCR), followed by a resistor and a single TV damper diode - 6AU4.
This combo has less voltage drop then the above, so I have to re-adjust the RC component in the first stage of filtering to get a proper voltage. (Well, actually I didn't get it settled at the voltage precisely as before and still a little bit higher. Short of proper components on hand and time... )
Anyway it works OK this way. I really like the slow turn-on of the IH diode tube, and the damper diode seems even slower.
It was tested on table for about half an hour last night and started playing in the morning now for about an hour. I hate to say that, but it does sound different from the previous setup. The sound of full SS bridge + damper diode is a little bit 'stiffer' and slightly emphasized in higher mid. The previous half SS/half tube hybrid bridge sounds more laidback by comparison.
I thought the additional series resistor after the SS rectifier and RCLCLC filter shuold make all AC "issues" disappear, but it seems not entirely the case. The operation point of the amp is shifted a little, but only a few volts and several mA, should be no big deal.
Is it the snubber not properly set?
CLS,
Just for interest,
UF diodes will give a brighter sound + higher voltage.
Soft recovery like BYV96E will give you slightly less voltage and the "laidback" sound you talk about.
HEX are sort of in between because they are efficient like the UF diodes however they give a soft recovery!
The soft recovery seems to reduce the ringing at the expense of the lost volts. Probably because it takes longer to switch off!
All just for fun!
Regards
M. Gregg
Just for interest,
UF diodes will give a brighter sound + higher voltage.
Soft recovery like BYV96E will give you slightly less voltage and the "laidback" sound you talk about.
HEX are sort of in between because they are efficient like the UF diodes however they give a soft recovery!
The soft recovery seems to reduce the ringing at the expense of the lost volts. Probably because it takes longer to switch off!
All just for fun!
Regards
M. Gregg
Hi Gregg,
Thanks for the tips.
Later I tuned the snubbers a little and the sound did change. However I'm shooting in the dark here. First I'm not good at calcuating numbers. Second I don't know and can't measure the leakage inductances of the windings etc. So I'm short of the basic 'facts' to calculate. (I used to own an oscilloscope, but now it's stored at somewhere else... argh.... )
Now there're 2 RC, 1 C, 1 MOV across the HT secondary: 0.1u+12R // 0.01u+55R // 0.01u // 820V MOV (the later 3 are tied at the bridge).
And for filament supplies, there're 2200p+175R on the secondaries and 0.01u across each diode of the bridge.
At this stage of fiddling, it sounds alright, the egdiness is rounded. It was playing all day long yesterday and I was quite comfortable with the sound -- at least I let it play for that long and did not mod it again.
Or I should let the parts burn-in well and settle down... duh.
Thanks for the tips.
Later I tuned the snubbers a little and the sound did change. However I'm shooting in the dark here. First I'm not good at calcuating numbers. Second I don't know and can't measure the leakage inductances of the windings etc. So I'm short of the basic 'facts' to calculate. (I used to own an oscilloscope, but now it's stored at somewhere else...
argh.... )Now there're 2 RC, 1 C, 1 MOV across the HT secondary: 0.1u+12R // 0.01u+55R // 0.01u // 820V MOV (the later 3 are tied at the bridge).
And for filament supplies, there're 2200p+175R on the secondaries and 0.01u across each diode of the bridge.
At this stage of fiddling, it sounds alright, the egdiness is rounded. It was playing all day long yesterday and I was quite comfortable with the sound -- at least I let it play for that long and did not mod it again.
Or I should let the parts burn-in well and settle down... duh.
CLS,
"Later I tuned the snubbers a little and the sound did change. However I'm shooting in the dark here."
Sometimes its interesting to try "tuning by ear" and find the sound that you like! Measurements are not the whole story!
You could build an amp that measures perfect, and when you listen it is a real let down!
My advise is to listen to it for a few weeks, also listen to other peoples equipment. If you find yours sounds "better to you" then you have your answer!
You may find if you change the diodes to BYV96E (soft recovery) you may miss the brighter sound of UF diodes.
Remember as you "improve the noise floor" of the amp it will make any other weaknesses in your system and recordings show!
Have you tried cleaning (removing transients) on the incoming mains on the primary of the power Tx. The transients appear on the peak of the AC and "show" during the switch off time of the diodes!
A few things just for fun!
Regards
M. Gregg
"Later I tuned the snubbers a little and the sound did change. However I'm shooting in the dark here."
Sometimes its interesting to try "tuning by ear" and find the sound that you like! Measurements are not the whole story!
You could build an amp that measures perfect, and when you listen it is a real let down!
My advise is to listen to it for a few weeks, also listen to other peoples equipment. If you find yours sounds "better to you" then you have your answer!
You may find if you change the diodes to BYV96E (soft recovery) you may miss the brighter sound of UF diodes.
Remember as you "improve the noise floor" of the amp it will make any other weaknesses in your system and recordings show!
Have you tried cleaning (removing transients) on the incoming mains on the primary of the power Tx. The transients appear on the peak of the AC and "show" during the switch off time of the diodes!
A few things just for fun!
Regards
M. Gregg
Have you tried cleaning (removing transients) on the incoming mains on the primary of the power Tx. The transients appear on the peak of the AC and "show" during the switch off time of the diodes!
How might you do this M? A couple of small caps across the rails separated by a couple of resistors?
CLS,
If you want to mod and try different things! You have a hole in your chassis from the Valve rectifier GZ34. You could fit an octal base and connect your connections to this. Then buy/make a slim box about the size of a tube and fit your mods in this! You can then plug and play different mods and compare without stripping the amp every few days!
Remember to be carefull that your HT will rise when you use more efficient diodes UF, and drop with soft recovery or tube rectifiers. I am thinking about your HT caps working voltage on power on with no heaters!
Just for fun.
Regards
M. Gregg
If you want to mod and try different things! You have a hole in your chassis from the Valve rectifier GZ34. You could fit an octal base and connect your connections to this. Then buy/make a slim box about the size of a tube and fit your mods in this! You can then plug and play different mods and compare without stripping the amp every few days!
Remember to be carefull that your HT will rise when you use more efficient diodes UF, and drop with soft recovery or tube rectifiers. I am thinking about your HT caps working voltage on power on with no heaters!
Just for fun.
Regards
M. Gregg
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.