Your solution of Allen's RTP3 preamp
Hi Henning,
Have also found now here the pictures of your really impressive realisation of the RTP3. Fantastic work, really awesome what you have made.
Although I have to say that I find it a pity that you didn't incorporate the phono stage since this is less complicated to construct than the line stage and vinyl reproduction has always been one of the hallmarks of Allen's design since he was always a big supporter of analog over digital audio.
But ok.
I have also two questions.
Looking at the top view of the power supply with so many inductors and other things I noticed that you did not follow Allen's RTP3D power supply design but made use of a quite different circuit design or? Can you tell anything about that?
Other questions is regarding the use of the LT3086 for the filaments of the tubes.
Any particular reason why you used these? According to the datasheet these are low dropout voltage regulators. So did you make individual voltage regulators for each tube?
I'm asking because Allen used LM317 regulators however these are not applied as voltage regulators but they are connected as current sources which Allen sonically preferred over voltage regulators as also mentioned in his TPCB.
Hope to hear from you here.
Thanks
Günter
Hi Henning,
Have also found now here the pictures of your really impressive realisation of the RTP3. Fantastic work, really awesome what you have made.
Although I have to say that I find it a pity that you didn't incorporate the phono stage since this is less complicated to construct than the line stage and vinyl reproduction has always been one of the hallmarks of Allen's design since he was always a big supporter of analog over digital audio.
But ok.
I have also two questions.
Looking at the top view of the power supply with so many inductors and other things I noticed that you did not follow Allen's RTP3D power supply design but made use of a quite different circuit design or? Can you tell anything about that?
Other questions is regarding the use of the LT3086 for the filaments of the tubes.
Any particular reason why you used these? According to the datasheet these are low dropout voltage regulators. So did you make individual voltage regulators for each tube?
I'm asking because Allen used LM317 regulators however these are not applied as voltage regulators but they are connected as current sources which Allen sonically preferred over voltage regulators as also mentioned in his TPCB.
Hope to hear from you here.
Thanks
Günter
Hi Günter,
many thanks for your kind words! And indeed, it was a tough decision for me to omit the phono stage in my build. But being in my late 30's, I have not grown up with phono and don't own any vinyl.
The power supply is a complete redesign, because I could gain a certain amount of practical know-how around good PSU-designs in the past.
I've realised the following voltages:
* in total three independent, to 7,2V preregulated, heater supplies (1x with 60V above GND, shared with left and right channel and 2x with 175V above GND for upper tubes, one per channel)
* 36V preregulated via self-designed Sigma11 regulator
* -34V preregulated via self-designed Sigma11 regulator
* 12V for remote power switching (the power switch on the front of the RPT3D triggers mosfet-relais for the low voltages and delay-timers for the HV-Supply in the PSU
* 350V raw voltage for the B+ shunt regulator
All small voltages are "actively" rectified based on LT4320, all except 12V feature an input choke filter and Jensen 4-pole electrolytic caps as main filter Cap.
The raw B+ supply is filtered via RcLCLC network to 4mV remaining ripple under load according PSUD2 simulation.
My choice to use voltage regulators instead of current sources for powering the heaters was based on practical consideration. I just wanted to be free what breed of E88CC tube I would use in future and didn't want to have any hassle about feeding the heaters with the wrong current.
I've chosen LT3086 because of their low drop voltage on one hand (to safe some heat inside the RTP3D chassis) and - more importantly - to make use of the current limit feature, which I set to 444mA to ensure a soft start of cold tubes.
Best regards
Henning
many thanks for your kind words! And indeed, it was a tough decision for me to omit the phono stage in my build. But being in my late 30's, I have not grown up with phono and don't own any vinyl.
The power supply is a complete redesign, because I could gain a certain amount of practical know-how around good PSU-designs in the past.
I've realised the following voltages:
* in total three independent, to 7,2V preregulated, heater supplies (1x with 60V above GND, shared with left and right channel and 2x with 175V above GND for upper tubes, one per channel)
* 36V preregulated via self-designed Sigma11 regulator
* -34V preregulated via self-designed Sigma11 regulator
* 12V for remote power switching (the power switch on the front of the RPT3D triggers mosfet-relais for the low voltages and delay-timers for the HV-Supply in the PSU
* 350V raw voltage for the B+ shunt regulator
All small voltages are "actively" rectified based on LT4320, all except 12V feature an input choke filter and Jensen 4-pole electrolytic caps as main filter Cap.
The raw B+ supply is filtered via RcLCLC network to 4mV remaining ripple under load according PSUD2 simulation.
My choice to use voltage regulators instead of current sources for powering the heaters was based on practical consideration. I just wanted to be free what breed of E88CC tube I would use in future and didn't want to have any hassle about feeding the heaters with the wrong current.
I've chosen LT3086 because of their low drop voltage on one hand (to safe some heat inside the RTP3D chassis) and - more importantly - to make use of the current limit feature, which I set to 444mA to ensure a soft start of cold tubes.
Best regards
Henning
Hi all,
here is a noise plot, taken from the output of my RTP3D build with a Focusrite Forte audio interface. Input was not shorted.
In the meantime I could get rid of the small humps at 50 and 100Hz after optimising the Ground connection between the PSU und the Preamp.
-140dBV is the noisefloor of the audio interface, the spike at 20kHz is an artefact of the ADC.
Best, Henning
here is a noise plot, taken from the output of my RTP3D build with a Focusrite Forte audio interface. Input was not shorted.
In the meantime I could get rid of the small humps at 50 and 100Hz after optimising the Ground connection between the PSU und the Preamp.
-140dBV is the noisefloor of the audio interface, the spike at 20kHz is an artefact of the ADC.
Best, Henning
I am not an expert, but I dare to state that the standard is voltage regulation and allowing current to do its things, except when current limited. Voltage is pressure and amp is flow, so you control the pressure and the flow is what it is.
What I found odd, just a bit, is the notion of "better" because one control current and not the voltage. What is better anyway, did he explain that or are we dealing with a biased perspective ?
Take the triode, almost everyone talk about B+ voltage, not B+ current and you can tune the performance of the triode by using different B+ voltages. At least does who have built multiple tube amps report a difference in sonic behavior when B+ voltage is adjusted up or down.
And besides, like HenSch, I would like to try different tubes, 6H30, 6N6P, 6E5P or 6N5P.
What I found odd, just a bit, is the notion of "better" because one control current and not the voltage. What is better anyway, did he explain that or are we dealing with a biased perspective ?
Take the triode, almost everyone talk about B+ voltage, not B+ current and you can tune the performance of the triode by using different B+ voltages. At least does who have built multiple tube amps report a difference in sonic behavior when B+ voltage is adjusted up or down.
And besides, like HenSch, I would like to try different tubes, 6H30, 6N6P, 6E5P or 6N5P.
Oneminde,
You are absolutely right that voltage regulation is the standard for filament power supply but that does not necessarily mean that it is the best solution from a sonical point of view and that is what I said about Allen's choice to apply current sources for the filaments.
Here is what he said in his TPCB:
"While on the heater topic, a good tweak is to feed each tube from it's own LM317 constant current source. This gives a controlled warm up to save those special tubes and a quieter background and better stage to stage isolation than using a voltage regulator".
Of course a voltage regulator is more convenient to use when it comes to feed ECC88 kind of tubes and other 6,3 V heater tubes like the ones you mentioned since you simply set the regulator to 6.3 V and then it will sort out the required current.
With a current source you will have to check i.e. make it adjustable to set the current for reaching the required 6.3 V for each tube individually but it can be done and I personally have no problem with that.
By the way for PCC88 kind of tubes which are used right now in the commercial RTP3's as far as I know it is definitely better to apply current sources because their filaments are current and not voltage driven.
The comparison with B+ voltages makes not so much sense in my opinion since much lower currents are involved and we are talking there about high and not low voltages.
But even there you can find tube circuit concepts where the anode ( or cathode ) resistor is replaced with a constant current source independent if and how B+ is regulated or not.
Allen applied this in the cathodes of all stages in the RTP3 for good reasons as also described in his TPCB.
Best regards,
Günter
You are absolutely right that voltage regulation is the standard for filament power supply but that does not necessarily mean that it is the best solution from a sonical point of view and that is what I said about Allen's choice to apply current sources for the filaments.
Here is what he said in his TPCB:
"While on the heater topic, a good tweak is to feed each tube from it's own LM317 constant current source. This gives a controlled warm up to save those special tubes and a quieter background and better stage to stage isolation than using a voltage regulator".
Of course a voltage regulator is more convenient to use when it comes to feed ECC88 kind of tubes and other 6,3 V heater tubes like the ones you mentioned since you simply set the regulator to 6.3 V and then it will sort out the required current.
With a current source you will have to check i.e. make it adjustable to set the current for reaching the required 6.3 V for each tube individually but it can be done and I personally have no problem with that.
By the way for PCC88 kind of tubes which are used right now in the commercial RTP3's as far as I know it is definitely better to apply current sources because their filaments are current and not voltage driven.
The comparison with B+ voltages makes not so much sense in my opinion since much lower currents are involved and we are talking there about high and not low voltages.
But even there you can find tube circuit concepts where the anode ( or cathode ) resistor is replaced with a constant current source independent if and how B+ is regulated or not.
Allen applied this in the cathodes of all stages in the RTP3 for good reasons as also described in his TPCB.
Best regards,
Günter
Henning,
Thank you very much for answering my questions.
Interesting that you come to a completely different solution for the B+ supply with a
RCLCLC topology whereas Allen applied LCRCRC in the RTP3D.
On the heaters.
Ok, I can understand now why you didn't want to use current sources.
Don't know which tubes you are using right now but when you limit the heater current to 444 mA two of the most interesting tubes possible to apply here in the gain stage cannot be used.
One is the E288CC ( in my opinion sonically clearly better than std. xx88 tubes ) with much higher transconductance ( 20ma/V vs 12,5 ma/V ) but with lower gain and the other is the well known 6H30 tube also with lower gain and 18mA/V transconductance.
Both tubes require higher heater currents ( 474 mA for the E288CC and even 850 mA for the 6H30 ).
The E288CC is a direct replacement so no other changes necessary whereas a 6H30 would require more changes due to different pin configuration and probably adaptation of operating point.
Both tubes are larger in size than std. xx88 tubes so operating these tubes may
require removing the top cover.
Nevertheless worth a try I think.
Best regards,
Günter
Thank you very much for answering my questions.
Interesting that you come to a completely different solution for the B+ supply with a
RCLCLC topology whereas Allen applied LCRCRC in the RTP3D.
On the heaters.
Ok, I can understand now why you didn't want to use current sources.
Don't know which tubes you are using right now but when you limit the heater current to 444 mA two of the most interesting tubes possible to apply here in the gain stage cannot be used.
One is the E288CC ( in my opinion sonically clearly better than std. xx88 tubes ) with much higher transconductance ( 20ma/V vs 12,5 ma/V ) but with lower gain and the other is the well known 6H30 tube also with lower gain and 18mA/V transconductance.
Both tubes require higher heater currents ( 474 mA for the E288CC and even 850 mA for the 6H30 ).
The E288CC is a direct replacement so no other changes necessary whereas a 6H30 would require more changes due to different pin configuration and probably adaptation of operating point.
Both tubes are larger in size than std. xx88 tubes so operating these tubes may
require removing the top cover.
Nevertheless worth a try I think.
Best regards,
Günter
Regarding tube selection. How detailed or clean the tube sound is somewhat connected to Miller capacitance, and generally speaking: The lower the capacitance the cleaner and less distorted the signal will be.
Attachments
Oneminde,
Interesting your remark on Miller Capacitance and its impact on sonics although I'm not sure if this impact is true and major for all kinds of tube circuits.
For cascodes like applied in the RTP3 gain stages Allen made a remark, I think I have read that in Audioasylum, that transconductance of a tube is the most important factor for operation in a cascode circuit.
Regards,
Günter
Interesting your remark on Miller Capacitance and its impact on sonics although I'm not sure if this impact is true and major for all kinds of tube circuits.
For cascodes like applied in the RTP3 gain stages Allen made a remark, I think I have read that in Audioasylum, that transconductance of a tube is the most important factor for operation in a cascode circuit.
Regards,
Günter
Hi,
Yes, here a remark from Allen on cascode tube choice in Audioasylum although he forgot here to mention the E288CC:
RE: Allen - would your cascode work with DHTs? - Allen Wright - Tube DIY Asylum
On input capacitance he said the following in the schematic explanation of the line stage cascode in the instructions:
"The cascode topology also greatly reduces the effective input capacitance
of the first tube—allowing a bandwidth beyond 1MHz, which is little effected by the source impedance."
So this probably makes clear that input capacitance is less important for a cascode.
Regards,
Günter
Yes, here a remark from Allen on cascode tube choice in Audioasylum although he forgot here to mention the E288CC:
RE: Allen - would your cascode work with DHTs? - Allen Wright - Tube DIY Asylum
On input capacitance he said the following in the schematic explanation of the line stage cascode in the instructions:
"The cascode topology also greatly reduces the effective input capacitance
of the first tube—allowing a bandwidth beyond 1MHz, which is little effected by the source impedance."
So this probably makes clear that input capacitance is less important for a cascode.
Regards,
Günter
Right, for a phono stage it would be good. So if; "The cascode topology also greatly reduces the effective input capacitance of the first tube—allowing a bandwidth beyond 1MHz, which is little effected by the source impedance." - then the 6H30 will do since has higher transductance than 6N6P.
Good convo
Good convo
Just got around to finish my build, which is a combination of the RTP5 phono stage and the line stage from Acoustic Research AR25 - yes its heresy, don't ask
My project started in 2012 but in lockdown I was able to finish it, finally..... listening to a record right now. At first , it sounded like a cheap transistor radio, but a week later now and quite a bit of burn-in, its much much much better - lots of detail, deep soundstage and laser sharp positioning, frightening at some times....
I do have a little issue however, and wanted to ask if anybody here is experiencing the same thing - when comparing the volume to a CD player, the phono gain is way too low. Need to turn up volume all the way to get decent output, while CD stays at roughly 1/3 to 1/2. I then simulated the phono and in fact it is confirmed that the gain is lowish, 73dB at lot frequencies and around 57dB at 1kHz, where your regular MC phono stage should have something like 63-65dB, so I am lacking an extra 6-10dB.
Anybody has any idea what to do? I tried (in simulation though) multiple different input devices, like MAT12 or SSM2210, to no avail. I also increased the impedance of the RIAA network x2, ending up with a big drop in bass, from 200Hz downwards.
Would changing the tubes help at all (I have 6922EH in there)?
My project started in 2012 but in lockdown I was able to finish it, finally..... listening to a record right now. At first , it sounded like a cheap transistor radio, but a week later now and quite a bit of burn-in, its much much much better - lots of detail, deep soundstage and laser sharp positioning, frightening at some times....
I do have a little issue however, and wanted to ask if anybody here is experiencing the same thing - when comparing the volume to a CD player, the phono gain is way too low. Need to turn up volume all the way to get decent output, while CD stays at roughly 1/3 to 1/2. I then simulated the phono and in fact it is confirmed that the gain is lowish, 73dB at lot frequencies and around 57dB at 1kHz, where your regular MC phono stage should have something like 63-65dB, so I am lacking an extra 6-10dB.
Anybody has any idea what to do? I tried (in simulation though) multiple different input devices, like MAT12 or SSM2210, to no avail. I also increased the impedance of the RIAA network x2, ending up with a big drop in bass, from 200Hz downwards.
Would changing the tubes help at all (I have 6922EH in there)?
This is a "feature" of all of Allen's preamps! The front end gain is limited (the RTP3 in particular only has a single amplification stage there), so to get the total gain of the preamp high enough for a typical low-output MC cartridge, 50dB or so (at 1kHz) in the phono stage is combined with 10-20dB in the line stage to get up to the usual 60-70dB for an MC. This means that all the preamps end up having L-pad attenuators on the line inputs to get comparable gain for all inputs.
My RTP3C and SVP, both using 6922EHs with a hybrid phono input stage (the RTP3 uses SSM2210 at the input and the SVP has 2SK170), have a similar gain structure. I attenuate all the line stage inputs by 6-10dB.
Alex
My RTP3C and SVP, both using 6922EHs with a hybrid phono input stage (the RTP3 uses SSM2210 at the input and the SVP has 2SK170), have a similar gain structure. I attenuate all the line stage inputs by 6-10dB.
Alex