Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: kagliostro on February 12, 2016, 12:56:04 pm
-
Is some time I'm going around the idea of a 5C1 with a 1625 power tube (I have some one and they are cheaper than 807 for me)
Today is raining and I've no fancy to go on my shop to do something
so I decided to put together my ideas and I've draw this preliminary version of the schematic
(big trouble, I'm not confident with octal tubes components values, I tried to pilfer on various schematics on the web)
Every advice, comment, suggestion will be appreciated
Grazie
Franco
-
you will want SOME minimum resistance in the NFB loop. (10Kpot +15K R?) experiment.
tone stack is not grounded.
grid leak sucks: overloads easily. try 6SJ7 with matchless EF86 values: you will likely prefer them.
v2a has no bias or get rid of .02 CC to DC bias. if you don't want DC coupling add auto bias parts Rs. (1M + Rk). 22k Rk for V2a? are you sure?
--pete
-
If you want to save a couple of components you could try taking the voltage to 6SJ7 screen grid from the power tube cathode. I'm not sure if any commercial amp ever did this but there's good 20-30V there and that's exactly what the 6SJ7 wants if you're using a high value plate resistor.
-
Thanks Pete & Lauri
@ Pete, I tried to follow your indication, absolutely not sure about the catode resistor on the CF
@ Lauri, I'll try to find a bit of documentation about your council and to study it
Grazie
Franco
-
You still need a resistor between the cathode and the top of that 22K if you stay with that bootstrapped CF. Here's a much simpler solution that you can see in the Marshall Plexi. Just throw away that .02 coupling cap and the 1M resistor. Connect V1 plate directly to V2A grid. All done.
-
You still need a resistor between the cathode and the top of that 22K if you stay with that bootstrapped CF. Here's a much simpler solution that you can see in the Marshall Plexi. Just throw away that .02 coupling cap and the 1M resistor. Connect V1 plate directly to V2A grid. All done.
plate of 6sj7 may be too low to dc couple with proper bias @ 1/2 b+.
-
Thanks Pete & Lauri
@ Pete, I tried to follow your indication, absolutely not sure about the catode resistor on the CF
@ Lauri, I'll try to find a bit of documentation about your council and to study it
Grazie
Franco
add 1k-1.5k R between cathode and 22k.
-
Thanks Steve & Pete
I added the resistor on the catode of V1a, I must give a look around to perform a better knowledge on the CF
Grazie
Franco
-
You changed the first stage to cathode biased but forgot to remove the capacitor and 5M resistor from the control grid. Remove them and put a 1M resistor to ground before the 75k resistor.
-
Opps
Franco
-
i estimated 280V for B+ for CF. with a 1K bias R to balance CF at near 1/2B+ we need 100K tail. this puts the grid at about 130V. with 150V across the tube and the grid bias at -1.3V with respect to the cathode. if B+ remains constant, then increasing bias current raises the grid reference.
to balance at 140V/140V with 22K tail is not possible. the best you can hope for is 5mA of bias with no room for drift. we want bias at about 1-1.5V with the 6SL7, so build around that. you can make B+ what ever you like (within reason) and work from there.
suggest: 100K R tail, 1K Rbias, with 280V B+. should see 130V at Vg1 and 150V across tube. close enough.
--pete
-
Grazie Pete
I'll draw a new version of the schematic (hoping I've correctly understand your councils - my translation sometime is bad)
Franco
-
Well (or not) this is the schematic I've redrawn
Pete I tried to arrange V1 in a different way than the Matchless DC30 because on the notes I've on my archive there is something about the fact that pentodes set for high gain on V1 are prone to be microphonic, this was told for the EF86 on Vox and there the ra was 220k and rk 2.2k on the Matchless more, ra 330k and rk 2.2k (the same), so I've think to lower this values (I've seen it applied to an ecc83 for a gain of 120 with 250v B+)
The CF is completely revised, I borrowed it from an explanation about CF where the tube was an ecc82, there were two versions of the circuit, one without the 10uF cap and 10k resistor, the version with the 10uF cap is planned to solve any noise coming with B+ voltage
Definitively there is no more remembrance of the 5C1 here and I feel like someone on a small raft on the sea very far from the coast :sad: :sad:
I fear to lose myself and not get anything
I hope someone can help me, I really kicked into a quagmire
Franco
p.s.: Just curious, there are metal 6SL7 ?
-
Your cathode follower is no good. It has 10k input impedance and too high plate dissipation. Also the coupling cap after the cathode follower goes straight to ground.
-
Thanks Lauri
I put a wrong value to that resistor (the 10k resistor) because in the schematic I was watching the value was missing
may be a 100k resistor is correct there ???
the ground on the Tone Control is on the wrong side (I reversed the drawing)
here 2 new versions of the schematic
one with the CF bias circuit semplified
Franco
-
revision 05 of your schematic looks better but you still need to increase the load resistor of the cathode follower. With 10k resistor the tube is drawing quite a lot of current and the plate dissipation is about 1W.
The cathode follower circuit you had in your revision 02 is much better, just increase the bias resistor to 1.5k or 2.2k and increase the load resistor to 100k. Or if you want to use the cathode follower in revision 5 then just increase the load resistor to 56k-100k.
-
1.8m /470k divider: the 470k needs to be 1.8m. change 10k under v2a to 100k.
--pete
-
I looked at the tube data sheet for the 1625 power tube. It shows maximum plate voltage of 600 to 700 volts and maximum screen voltage of 250 volts. If I remember correctly from my Plexi build a bridge rectifier effectively doubles the voltage from the transformer so your 340 volts at 150mA would become 680 volts AC. When converted to DC would be even higher, maybe 900 volts. On top of that not sure how a 22K resistor would knock the screen voltage down to 300 volts which is 50 volts over the maximum per the datasheet.
Don't know if there is any validity to these comments but it is what came to mind when I looked at the schematic.
Thanks
Mike
-
no. FWB is 1.414 x Vrms. so 480v.
--pete
-
There was a problem exporting the schematics in .jpg format
here is the right 05 schematic (now 05bis)
are this councils to be applied ??
1.8m /470k divider: the 470k needs to be 1.8m. change 10k under v2a to 100k.
with 1.8M and 470k voltage divider, assuming B+ 300v I've ~62v on the grid
On my reference book the cathode resistor is in the order of 100k for DC coupled CF, for AC coupled CF the value is 10-47k
have you taken in consideration that this is an AC coupled CF ?
Thanks
Franco
-
no. FWB is 1.414 x Vrms. so 480v.
--pete
Thanks Pete, I missed the fact that the output from the bridge rectifier didn't go to the primary center tap of the OT. Need to make a file of different ways to wire a power supply.
Thanks
Mike
-
Mike what Pete is saying is that with 320v DC after the bridge on the first capacitor you'll find a DC voltage that is 340v x 1.414 = 480v
then there is an inductor where the voltage drop a bit so on node A the voltage will be less than the 480v of cours the voltage decrease on node B and C
Franco
-
with 1.8M and 470k voltage divider, assuming B+ 300v I've ~62v on the grid
62V on the grid means you also have about the same on the cathode. With 10k cathode resistor it means you are pulling over 6mA of current which means with 300V on the plate you're way over the maximum plate dissipation for 6SL7.
-
OK, so what to do ?
Lower the voltage on plate or change the value on the resistor divider or catode resistor value ???
Too difficult for me, as I told, I'm stick in the mud
Franco
-
Doing a search I've find an AC CF that uses ecc83 tube that has a plate dissipation of 1W like the 6SL7
I stolen the arrangement and components values and put it on this new version of the schematic
so I'm back to the suggestion Pete give me previously only with a red LED (~1.6v) instead of a resistor
Franco
-
OK, so what to do ?
Lower the voltage on plate or change the value on the resistor divider or catode resistor value ???
Too difficult for me, as I told, I'm stick in the mud
Franco
Increase the value of cathode resistor to decrease current.
Doing a search I've find an AC CF that uses ecc83 tube that has a plate dissipation of 1W like the 6SL7
I stolen the arrangement and components values and put it on this new version of the schematic
so I'm back to the suggestion Pete give me previously only with a red LED (~1.6v) instead of a resistor
Franco
470k resistor from grid should go to between the led and 47k resistor, not to ground.
-
see attached. there are two basic methods to AC couple a CF. the self biased (also called auto bias) and fixed bias. the objective is for the valve to sit at 1/2 B+ at the control grid. think of it as a virtual ground at 1/2 B+. with B+ at 280V, you have positive 1/2 of wave is 140V and negative half is also 140V.
in your plan with Vg sitting at 62V, the maximum un-distorted signal is +/-62V because virtual ground is at 62V, or Vg=62V.
for auto bias you need to calculate bias and Vg with load lines.
with fixed bias, 1/2B+/Rcath is bias current. so in this case plate dissipation is 1.4mA * 140V or ~ 200mW
--pete
-
> OK, so what to do ?
The tube DC load resistor is selected to drive the following AC load.
Here the load is the James tone network. At medium frequency this will approach 68K+27K= 95K.
(Note that "rev 5bis" shows the bass side of the tone network upside down; 68K goes to the top, input, side.)
For a cathode follower in a "small signal" stage, pick the DC resistor somewhat less to maybe 1/3rd of the load. 90K to 30K? 47K seems a good value.
There are many clever ways to bias cathode followers. I would strongly urge- Keep It Simple. Find the Resistance-Coupled Amplifier table for this tube. Find the line for 300V supply and a load similar to 95K. Use those plate and cathode resistor values, except in cathode-follower plan.
http://www.tubebooks.org/tubedata/hb-3/Receiving_Tubes_Part_1/Resistance-Coupled_Amplifiers.PDF (http://www.tubebooks.org/tubedata/hb-3/Receiving_Tubes_Part_1/Resistance-Coupled_Amplifiers.PDF)
6SL7, 100K load, use 100K "plate" resistor (but put it in the cathode return), 1500 Ohm cathode bias resistor. Maximum output can be 40V peak. Gain as CF will be about (VG/(1+VG)) or (34/(1+34)) or 0.97.
EDIT: just like in Pete's left-hand plan. Note: "Vg = 1/2 B+" may not be true, and can not be measured at the grid with any common meter. If the Cathode comes out in the range 1/3rd to 2/3rd of B+, you are fine. (But beware heater-cathode voltage limits.)
-
clarification: see attached please.
--pete
-
k, there are no metal 6SL7 only glass.
--pete
-
Excuse me if I answer only now
I was a bit engaged on my other hobby (shooting), a friend asked me to try to repair this big trouble on an old Anshutz 1408 Match 54 22LR stock
(http://i.imgur.com/TfgfHPI.jpg)
and yesterday I was at the Hit Show on Vicenza (very nice showcase)
https://www.all4shooters.com/it/Tiro/news-generiche/eventi/HIT-Show-apertura-iscrizioni-espositori/ (https://www.all4shooters.com/it/Tiro/news-generiche/eventi/HIT-Show-apertura-iscrizioni-espositori/)
Thanks PRR & Pete
I'm a bit confused ....
For a cathode follower in a "small signal" stage, pick the DC resistor somewhat less to maybe 1/3rd of the load. 90K to 30K? 47K seems a good value.
............ 6SL7, 100K load, use 100K "plate" resistor (but put it in the cathode return), 1500 Ohm cathode bias resistor.
Here attached is a new version of the schematic (hoping I've correcltly understand you)
Considerations and councils about this 07 version ? There are changes to apply to the whole "project" ?
Grazie
Franco
-
grid leak is too high for 807 (1625) 1Mohm should be 500K or less with cathode bias.
http://w5jgv.com/downloads/RCA%20807.pdf (http://w5jgv.com/downloads/RCA%20807.pdf)
also, if wiper fails...so does tube and OT if you're not fast enough on the power switch.
IMO, that's a bad place for MV. replace 1M pot w/ 220K - 470K fixed R. if you want to leave it there, then at least use 500K pot and a 2.2M fixed R from wiper to gnd.
suggest moving it between V1 and V2 CF. if you do, see attached for connection schema.
--pete
-
Ciao Pete
Thanks for indications
What will happen if I reverse the connection of the MV ?? (rev 08)
EDIT:
The rev 09 is the schematic following your suggestions
The two resistors in red color are 560K in your schematic, is 560K the value to be used or is due to a copy and paste fault ?
Thanks
Franco
-
kagliostro, did you complete this build and have a final schematic? If so is it in the archives?
-
Due to healt problems of my father I stopped the project, then I swapped to a different plan using the same tubes
an AA764 with a second channel that can be used as an OD
(http://i.imgur.com/YRgGDgR.jpg)
(http://i.imgur.com/5Po3HIM.jpg)
(http://i.imgur.com/eJQaxRG.jpg)
Franco
-
It is my understanding that the lowly 6L6 and the 807 have similar characteristics. One of the design differences is plate wiring, where the 6L6 plate is hooked up at the socket, rather than having a separate wire at the top of the tube.
I understand your desire to keep your amp maintenance at low so I can see why you are looking at 1625 in place of the 807. 'nuff said
-
1625 is a 12.6v heater version of the 807 and 807 is a TopCap version of the 6L6 (or may be the 6L6 is a without TopCap version of the 807)
I can get 1625 tubes at 5€ each one (till now)
Franco
p.s.: The first mosfet on PS (near node A) is a fake inductor (or something that acts near an inductor respect the ripple)