I have been very interested by Merlin book on Tube Preamps and I decided to build the RIAA amplifier, however I decided to replace the cathode follower in the output stage by a MOSFET source follower (I already used this solution in a Baby Huey amplifier with a very good result) to save one tube and because I wanted to avoid the big voltage difference between the heater and the cathode of the output tube...
As I don't like point to point soldering, I have started to design a PCB. If someone has any suggestion about this project, please let me know.
Do you know if the PC900 can also be a good device for this project ? The pin-out is not compatible however 😡
Cheers,
Marc
As I don't like point to point soldering, I have started to design a PCB. If someone has any suggestion about this project, please let me know.
Do you know if the PC900 can also be a good device for this project ? The pin-out is not compatible however 😡
Cheers,
Marc
I don't understand the function of R12 (500R) trimmer.
If is it do balancing, it can work (IMHO it is unnecessary), but increasing of C6 blocking capacitor's ESR with part of R12 is the worst scenario.
If is it do balancing, it can work (IMHO it is unnecessary), but increasing of C6 blocking capacitor's ESR with part of R12 is the worst scenario.
I would suggest you alter D1 and R15 to follow my recommendation in fig. 7.32 of the book.
Hi Ian,
Did you suggest to use a current source like the ring of two used in the Baby Huey?
Hi Merlin,
Thanks for the suggestion, I made the change for R15, but I don't understand if D1 is OK like that on the new schematic ?
I have also added two 6.8 ohms resistors in serial with the heaters of the PC97 to reduce the 6.3V to the 4.5V 300 mA required...
Here are a new schematic with an improved PCB.
Cheers,
Marc
Did you suggest to use a current source like the ring of two used in the Baby Huey?
Hi Merlin,
Thanks for the suggestion, I made the change for R15, but I don't understand if D1 is OK like that on the new schematic ?
I have also added two 6.8 ohms resistors in serial with the heaters of the PC97 to reduce the 6.3V to the 4.5V 300 mA required...
Here are a new schematic with an improved PCB.
Cheers,
Marc
An interesting project. I could build one too ....
A suggestion : why not use some board space for a 7805 to create the 6.3VDC , thus
only some 9-15V DC is needed and the regulated filement is contained in a small area.
A suggestion : why not use some board space for a 7805 to create the 6.3VDC , thus
only some 9-15V DC is needed and the regulated filement is contained in a small area.
In a side question to merlin : what has the PC97 that cannot be done with an ecc83 or 88 ?
The book gives no clue of any specific reason other then the mere existence of some tubes.
Or did i miss some here ?
The book gives no clue of any specific reason other then the mere existence of some tubes.
Or did i miss some here ?
I'm not Merlin, but since I choose tubes for somewhat similar reasons, I'll have a go at it:
Low cost for NOS high quality tubes because of the odd filament voltage no one outside of the hobby sector wants them, comparatively high mu, high transconductance, and low rp.
ECC88 has low rp, but mu is not comparable.
ECC83 has high mu, low transconductance and high rp.
High transconductance and low rp are major benefits at any given level of mu. (better noise performance, etc.)
Low cost for NOS high quality tubes because of the odd filament voltage no one outside of the hobby sector wants them, comparatively high mu, high transconductance, and low rp.
ECC88 has low rp, but mu is not comparable.
ECC83 has high mu, low transconductance and high rp.
High transconductance and low rp are major benefits at any given level of mu. (better noise performance, etc.)
In his book, Merlin explain that the PC97 has a very low "grid-anode capacitance of just 0.5 pF" and "its Miller capacitance is low at 56 pF, so the external stray capacitance must only be about 0.8 pF" (page 354).
I was also thinking to use the PC900 which is the successor of the PC97 and seem to have even lower Cag @ 0.365 pF !
I have added the two datasheet.
I was also thinking to use the PC900 which is the successor of the PC97 and seem to have even lower Cag @ 0.365 pF !
I have added the two datasheet.
Yes that is perfect!
I would not bother with a CCS for the source follower as it is already highly linear and is expected to drive fairly heavy loads (10k) which largely negates the high impedance of a CCS.
Thanks for the feedback Merlin 🙂
What do you think about the PC900 tube ?
While I was trying to find a way to accommodate both tubes on the PCB ( they are not fully compatible, pin 1 is the grid instead of the cathode on the PC900) I discovered that my socket layout was wrong and I changed it...
I have also found a US equivalent 6FY5 (except for the 6.3V heater voltage).
What do you think about the PC900 tube ?
While I was trying to find a way to accommodate both tubes on the PCB ( they are not fully compatible, pin 1 is the grid instead of the cathode on the PC900) I discovered that my socket layout was wrong and I changed it...
I have also found a US equivalent 6FY5 (except for the 6.3V heater voltage).
Supposedly it's an improved PC97, so you'd think it would be just as good. But improved for VHF doesn't mean the same as improved for AF, so who knows. A triode is a triode, at the end of the day, and you can't make such fine predictions of performance, you have to actually try it. I expect it will work though, provided it's not terribly microphonic.
Last edited:
start up?
With cathode follower the cathode resistor charges gradually on start up allowing the coupling capacitor to charge at low current, with no output . With source follower, you have immediately maximum high voltage as the precedent triode did not start draining. Even you limit the output voltage with a zener diode protecting the next stage you still need to delay linking it.
With cathode follower the cathode resistor charges gradually on start up allowing the coupling capacitor to charge at low current, with no output . With source follower, you have immediately maximum high voltage as the precedent triode did not start draining. Even you limit the output voltage with a zener diode protecting the next stage you still need to delay linking it.
I hope this will be the final version, checked and rechecked, but if you find any error please let me know before I send the files to the PCB manufacturer🙂
I have added a jumper to accept both the PC97 and the PC900, I use only pin 7 for the cathode and I can select pin 2 (PC97) or pin 1 (PC900) for the grid. Finally I didn't add the current source recommended by Ian because Merlin didn't found it necessary and mainly because I didn't want to make the PCB bigger 😀
To make it more readable I have attached top and bottom PCB separately. Again thanks for all the good suggestions...
Cheers,
Marc
I have added a jumper to accept both the PC97 and the PC900, I use only pin 7 for the cathode and I can select pin 2 (PC97) or pin 1 (PC900) for the grid. Finally I didn't add the current source recommended by Ian because Merlin didn't found it necessary and mainly because I didn't want to make the PCB bigger 😀
To make it more readable I have attached top and bottom PCB separately. Again thanks for all the good suggestions...
Cheers,
Marc
I would be uncomfortable with leaving pin-1 of the PC97 unconnected, owing to a comment I make on the bottom of page 240. Apart from that I see no problems with your schematic.
I agree with you that pin 1 (and pin 2 in the case of the PC900) unconnected is not very elegant but I was thinking that this is only critical at very high frequency (the tube was designed for that !) and I didn't want to add more jumper on the board 🙂 I have read your comments in the book and if needed I can always put a little wire between pin 1 and pin 7 😱
I will check the PCB one more time and send it the the subcontractor next week !
I will check the PCB one more time and send it the the subcontractor next week !
Hi Marc
I would suggest you measure the capacitance of your phono cables before using C21/C22.
Actual miller capacitance of high mu triodes is the achilles heel of this approach for MM phono. Check figure 10.13 in Merlin's book.
For PC900
Cag = 0.365 + 0.7stray = 1.1pF
Cgk = 3.3 + 0.7stray = 4pF
Amplification Factor around 70 (?)
Cmiller = 4pF + (70+1)*1.1pF = around 85pF
So you need to get the phono cables under 100pF. From my experience this can be challenging. Don't forget to add in the capacitance of your rca or xlr connector...
Ian
I would suggest you measure the capacitance of your phono cables before using C21/C22.
Actual miller capacitance of high mu triodes is the achilles heel of this approach for MM phono. Check figure 10.13 in Merlin's book.
For PC900
Cag = 0.365 + 0.7stray = 1.1pF
Cgk = 3.3 + 0.7stray = 4pF
Amplification Factor around 70 (?)
Cmiller = 4pF + (70+1)*1.1pF = around 85pF
So you need to get the phono cables under 100pF. From my experience this can be challenging. Don't forget to add in the capacitance of your rca or xlr connector...
Ian
Last edited:
Hi Ian,
Thanks for your feedback. I fully agree with your analysis, in fact I even didn't put these capacitors in my first version and it's only after reading Merlin comments at the end of page 358 that I decided to add this option (I had to redraw a part of my PCB to put them as close as possible of the input). With RCA connectors not on the PCB, I would suggest to solder them directly on the connector !
I have also considered to use a FET on the input as described on fig. 5.23 but finally I decided to make it full tube except of course for the MOSFET follower which have no gain and didn't change the audio tube sound.
Marc
Thanks for your feedback. I fully agree with your analysis, in fact I even didn't put these capacitors in my first version and it's only after reading Merlin comments at the end of page 358 that I decided to add this option (I had to redraw a part of my PCB to put them as close as possible of the input). With RCA connectors not on the PCB, I would suggest to solder them directly on the connector !
I have also considered to use a FET on the input as described on fig. 5.23 but finally I decided to make it full tube except of course for the MOSFET follower which have no gain and didn't change the audio tube sound.
Marc