Sebastian,
Look at the triode wired "El Cheapo" schematic. While it's only 6 WPC, it provides us with a point of departure. We will exploit a fact that Fisher, Scott, and Sherwood exploited, which is that "12" W. O/P tube driver circuitry also drives 7591s.
As I stated previously, you will build an "El Cheapo Grande".
Edcor's CXPP60-MS-6.6K O/P transformer has the correct primary impedance and the requisite magnetic headroom.
If your acquaintance can custom wind a power transformer and filter choke, things are looking up.
BTW, even though it's opamp based, Jim Hagerman's Bugle2 phono preamp could be the answer to your prayers.
Look at the triode wired "El Cheapo" schematic. While it's only 6 WPC, it provides us with a point of departure. We will exploit a fact that Fisher, Scott, and Sherwood exploited, which is that "12" W. O/P tube driver circuitry also drives 7591s.
As I stated previously, you will build an "El Cheapo Grande".Edcor's CXPP60-MS-6.6K O/P transformer has the correct primary impedance and the requisite magnetic headroom.
If your acquaintance can custom wind a power transformer and filter choke, things are looking up.
BTW, even though it's opamp based, Jim Hagerman's Bugle2 phono preamp could be the answer to your prayers.
Attachments
The only thing left to answer for is this question: The tannoy has a recommended amplification of minimum 10W. won't it be damaged if I run it with 6W?
Also, as stated above my friend does not wind transformers, he makes chassis, so I'll have to use Edcor transformers
Edit: Aaah I see. I can change the tubes to 7591s to increase output power?
Also, as stated above my friend does not wind transformers, he makes chassis, so I'll have to use Edcor transformers
Edit: Aaah I see. I can change the tubes to 7591s to increase output power?
Last edited:
Sebastian,
The schematic I uploaded is for a 6 WPC amp, based on "12" W. O/P tubes. You will build a 30 WPC amp, using the same topology style. Remember what I said about the 7591 being easy to drive. The 12AT7/ECC81 long tailed pair (LTP) has the "stones" needed to drive PP 7591s.
Definitely Edcor O/P transformers, with an agent in Germany. Maybe some money can be saved by ordering power "iron" from AnTek. While shipping from the USA is expensive, New Jersey is east coast and over 3000 KM. closer to Europe than New Mexico. It does matter.
AnTek's $38.50 AS-2T300 looks like it will do quite nicely. You might only have to buy a filter choke, in addition to what's already been mentioned. The AS-2T300 has a pair of 4 A. filament windings and the draw of 2X 12AT7/ECC81s and 4X 7591s is 3.8 A. So, the 2nd 4 A./6.3 VAC winding could be voltage multiplied to obtain the low current negative voltages (B- & C-) an "El Cheapo Grande" requires.
The schematic I uploaded is for a 6 WPC amp, based on "12" W. O/P tubes. You will build a 30 WPC amp, using the same topology style.
Remember what I said about the 7591 being easy to drive. The 12AT7/ECC81 long tailed pair (LTP) has the "stones" needed to drive PP 7591s. Definitely Edcor O/P transformers, with an agent in Germany. Maybe some money can be saved by ordering power "iron" from AnTek. While shipping from the USA is expensive, New Jersey is east coast and over 3000 KM. closer to Europe than New Mexico. It does matter.
AnTek's $38.50 AS-2T300 looks like it will do quite nicely. You might only have to buy a filter choke, in addition to what's already been mentioned. The AS-2T300 has a pair of 4 A. filament windings and the draw of 2X 12AT7/ECC81s and 4X 7591s is 3.8 A. So, the 2nd 4 A./6.3 VAC winding could be voltage multiplied to obtain the low current negative voltages (B- & C-) an "El Cheapo Grande" requires.
300B single ended amplifier - it is very simple, almost sure successful, easy to build... it has very warm and sweet sound...
Push-Pull is rather complicated that you would have to get the match pair of power output tube. Adjust on both AC and DC balancing.
Highly recommended 300B single ended.
Johnny
Push-Pull is rather complicated that you would have to get the match pair of power output tube. Adjust on both AC and DC balancing.
Highly recommended 300B single ended.
Johnny
Last edited:
30 watts per channel from a pair of 7591's implies pentode operation, correct? I don't think UL connection would get that much power from those, without going into class AB2 with source follower drivers. Or is that where you're thinking of going with this? Remember the Dynaco ST70 only really got 20W per channel, and that was with EL34's.
Speaking of EL34's... Why not run those at low plate volts (maybe 300V) and higher current (maybe 80mA per tube), deeeeep in class A. Maybe even in triode? They'd be easy to drive that way, and would be running at a low distortion operating point. You wouldn't get much power, though. JJ EL34's are OK, no?
--
Speaking of EL34's... Why not run those at low plate volts (maybe 300V) and higher current (maybe 80mA per tube), deeeeep in class A. Maybe even in triode? They'd be easy to drive that way, and would be running at a low distortion operating point. You wouldn't get much power, though. JJ EL34's are OK, no?
--
ALL of JJ's Octal production is extremely unreliable.
JJ knows how to make good tubes, but the will to do so is not present. 
We are going to run the 7591s in ultra-linear (UL) mode. The 7591 data sheet provides a sample PP, UL mode, "fixed" bias, operating condition set that yields 32 W. IMO, there's no reason we can't get close to that yield.
The 7591 is not rated for a positive control grid current regime. Therefore, Class "AB2" is out of consideration. 7591 gm is high, at 10.2 mA./V. The control grid helix has to be tight, in order to get that behavior. It seems there is no room for the radiation of heat generated by a positive grid current regime.
Last edited:
Sebastian,
The 300B is an expensive, directly heated, power triode. It's usually found in 8 WPC single ended (SE) amps. You will not get the power your speakers need and SE O/P transformers cost way more, for the power handled, than PP "iron" costs. Like a lot of us, you are financially constrained. Forget the 300B and concentrate on that which can be done, with the money available. Exercise care and you will achieve excellent results.
You really need to get technically up to speed, at least a little bit. Run a Google search for NEETS, which is U.S. Navy training material. NEETS is free, off the WWW.
What sort of test equipment do you have? A decent multi-meter is a must. More equipment, such as an oscilloscope, is nice, but "El Cheapo" style circuitry can be successfully executed, with only a good multi-meter on hand.
The 300B is an expensive, directly heated, power triode. It's usually found in 8 WPC single ended (SE) amps. You will not get the power your speakers need and SE O/P transformers cost way more, for the power handled, than PP "iron" costs. Like a lot of us, you are financially constrained. Forget the 300B and concentrate on that which can be done, with the money available. Exercise care and you will achieve excellent results.
You really need to get technically up to speed, at least a little bit. Run a Google search for NEETS, which is U.S. Navy training material. NEETS is free, off the WWW.
What sort of test equipment do you have? A decent multi-meter is a must. More equipment, such as an oscilloscope, is nice, but "El Cheapo" style circuitry can be successfully executed, with only a good multi-meter on hand.
S, take a look here; I have Mr. Lantino's VTA ST 120 (60/60 wpc) from a kit with all supplied. Here you can find lots of schematics etc. I like the higher power amps for all the headroom. Several members there from your area.
Free forum : The Dynaco Tube Audio Forum
Dale
Free forum : The Dynaco Tube Audio Forum
Dale
I'm reading the Tube part, really interesting, getting to know a lot by it, thanks Eli!
I'm getting to "understand" what you meant by driving 7591s but still, I would need a schematic. I am still reading through the amplifing part, and other stuff.
I am not completely dumb regarding electronics, I just didn't understood valves. I am working as a sound engineer, and learnt electronics, so I have the knowledge, but this NEET really helps me out!
I'm getting to "understand" what you meant by driving 7591s
but still, I would need a schematic. I am still reading through the amplifing part, and other stuff.I am not completely dumb regarding electronics, I just didn't understood valves. I am working as a sound engineer, and learnt electronics, so I have the knowledge, but this NEET really helps me out!
I took the liberty of throwing together an LTspice sim of a 12AT7 LTP driving PP 7591 UL into a Dyna A470 OPT (the OPT model I had handy) with an 8 ohm load on the 4 ohm tap (an attempt at a quick 'n dirty simulation of an 8k plate-plate load). The results are crazy good, even open loop, with no global feedback added. Probably too good to be true, as this is just a sim. But still, very promising.
I'll be following along with great interest.
RG
Sounds neat! Could you provide the schematic you used?
Gonna be soo happy if this turns out well, gonna provide every info you guys want if you help me
I don't want to publish a schematic that turns out to have mistakes. Also, this is Eli's idea, so I want to make sure it's OK with him. Eli?
RG
Sebestian,
Of course, Rongon can post what he simulated. The object of this exercise is a solidly performing amp, not aggrandizing my ego. The participation of other members can only assist in getting the job done.
You, naturally with some help, are going to draw the schematic that gets used. The signal portion of the "El Cheapo" schematic I previously uploaded will undergo only a few changes. The power supplies for this amp will be discussed in later posts.
Other than an increase in value of the 'T7 load resistors, which take advantage of the higher B+ rail voltage. The LTP splitter/driver stays "the same".
Current production EH 7591s are intolerant of violations of the published grid to ground to ground resistance limit and a reasonably large resistance in that position is needed, lest LTP gain become too small. That problem is solved by employing combination biased O/P tubes. "Stand" each pair of 7591s on a shared 100 Ω/470 μF. network (1 network in each channel). By making part of the total bias voltage self generated, we stabilize the operating conditions and acquire a convenient "idle" current set/test point. Additional benefits of the scheme are only a single bias trim pot. is needed in each channel and 7591 matching requirements are somewhat eased. Close matching of gm is always needed in a PP pair. However, as the shared RC bias network provides some protection against imbalance, minor variations in cathode current, between the 2 tubes of this PP pair, are OK.
Let's look at the circuitry's I/P. A 10 Kohm volume control is followed by a RC high pass filter. What the "corner" frequency of that filter is set to is governed by the capabilities of the O/P "iron". The Edcor stuff under consideration has plenty of magnetic headroom. Therefore, the "corner" freq. can safely be set below the audio band. Use a 0.068 μF. cap. to couple the pot. to the 'T7 grid circuitry. Use a 1% tolerance 140 Kohm metal film resistor in the grid to ground position. The caps. used in the 2 channels should be matched within 2%.
Of course, Rongon can post what he simulated. The object of this exercise is a solidly performing amp, not aggrandizing my ego. The participation of other members can only assist in getting the job done.
You, naturally with some help, are going to draw the schematic that gets used.
The signal portion of the "El Cheapo" schematic I previously uploaded will undergo only a few changes. The power supplies for this amp will be discussed in later posts.Other than an increase in value of the 'T7 load resistors, which take advantage of the higher B+ rail voltage. The LTP splitter/driver stays "the same".
Current production EH 7591s are intolerant of violations of the published grid to ground to ground resistance limit and a reasonably large resistance in that position is needed, lest LTP gain become too small. That problem is solved by employing combination biased O/P tubes.
"Stand" each pair of 7591s on a shared 100 Ω/470 μF. network (1 network in each channel). By making part of the total bias voltage self generated, we stabilize the operating conditions and acquire a convenient "idle" current set/test point. Additional benefits of the scheme are only a single bias trim pot. is needed in each channel and 7591 matching requirements are somewhat eased. Close matching of gm is always needed in a PP pair. However, as the shared RC bias network provides some protection against imbalance, minor variations in cathode current, between the 2 tubes of this PP pair, are OK.Let's look at the circuitry's I/P. A 10 Kohm volume control is followed by a RC high pass filter. What the "corner" frequency of that filter is set to is governed by the capabilities of the O/P "iron". The Edcor stuff under consideration has plenty of magnetic headroom. Therefore, the "corner" freq. can safely be set below the audio band. Use a 0.068 μF. cap. to couple the pot. to the 'T7 grid circuitry. Use a 1% tolerance 140 Kohm metal film resistor in the grid to ground position. The caps. used in the 2 channels should be matched within 2%.
As my understanding what you just said after a day of learning, I edited your schematic What you told what has to be done. Is this correct?
Of course 6.3V would be filament and Tube type is 7591, but looking forward to understand more!
One question crossed my mind: The 1 KΩ dale resistor is rated what power?
Of course 6.3V would be filament and Tube type is 7591, but looking forward to understand more!
One question crossed my mind: The 1 KΩ dale resistor is rated what power?
The 1 K part that connects the non-inverting triode's grid to ground carries very little current. A 1/2 W., 1%, metal film part is in order. My preference is for Vishay/Dale RN65 milspec stuff, but any decent metal film part is OK.
As this amp is going use UL mode "finals", the screen grid resistors of the O/P tubes connect to the O/P transformer's UL taps, not the tube anodes. 7591 screen grids are much tougher than 'AQ5 screen grids. So, reduce the value of the screen grid resistors to 100 Ω. Let's avoid a potential accident and use 1 W. carbon film parts here.
Remember, I said combination bias. The 470 K resistors become 330 K and, while they do tie together, they don't go directly to ground. Instead, they tie to the wiper of the "idle" current trim potentiometer.
High pass poles are ALWAYS a consideration. As the total grid to ground resistance is going down, the value of the interstage coupling caps. must go up. Use 0.15 μF. interstage coupling capacitors.
As this amp is going use UL mode "finals", the screen grid resistors of the O/P tubes connect to the O/P transformer's UL taps, not the tube anodes. 7591 screen grids are much tougher than 'AQ5 screen grids. So, reduce the value of the screen grid resistors to 100 Ω. Let's avoid a potential accident and use 1 W. carbon film parts here.
Remember, I said combination bias. The 470 K resistors become 330 K and, while they do tie together, they don't go directly to ground. Instead, they tie to the wiper of the "idle" current trim potentiometer.
High pass poles are ALWAYS a consideration. As the total grid to ground resistance is going down, the value of the interstage coupling caps. must go up. Use 0.15 μF. interstage coupling capacitors.
Okay, understood all you said, except the Bias part. which trim potentiometer should it connect to?
I'm drawing the schematic right now, and will post it as an update, Thank you for helping me Eli!
Also: where the double circle connect to In the LTP Cathode section? That is the circuit on the right side of the PSU, Using the B-?
Edit: Here it is! As I understood, this has to happen. Am I right? I could not make the bias part, as I did not understood the trim potentiometer part, so bare with me.
I'm drawing the schematic right now, and will post it as an update, Thank you for helping me Eli!
Also: where the double circle connect to In the LTP Cathode section? That is the circuit on the right side of the PSU, Using the B-?
Edit: Here it is! As I understood, this has to happen. Am I right? I could not make the bias part, as I did not understood the trim potentiometer part, so bare with me.
Last edited: