![]() ![]() |
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
24.27.41.140
In Reply to: RE: DHT 01a tube preamp posted by mwhouston on February 14, 2025 at 16:29:01
"To limit output I put a series resistor of 330K to the output RCA with 100K to ground at the RCA"Your later comments regarding HF response and testing indicate that you don't understand how this all works together. Below is the frequency response of the 300K/100K combination driving your 175pF cable. And this doesn't account for the additional output impedance of the 01A anode or input capacitance of the amplifier. In addition to high frequency loss, phase (dotted line) is badly shifted.
Edits: 02/16/25
Regarding the voltage divider . . . couldn't he just use a 33 ohm in series and a 10 ohm to ground? The voltage reduction would be the same.
The resulting output impedance would then be dominated by the impedance of the tube which, assuming a 50k load resistor, would be 8.33k.
If the formula I posted is correct (??) the output impedance would be:
(8.33k + 33 ohms) + (10 ohms || the assumed 100k input impedance of the amp) = 8363 ohms + 9.99 ohms = 8373 / 8.373k
Looks like this would move the rolloff point to ~108k if only the 175pf capacitance of the cable is included.
Am I missing something?
I built something similar to the Bartola but coupled the output with a TVC to reduce the output impedance (Thanks to Dave Slagle).
Its the best preamp I've had in 40 yrs, it will make a Krell SS amp sound good....
Edits: 02/17/25
The Bartola is taking the output from the mu output connection on the lower device of the CCS plate load. That works like a cathode follower and should have a pretty low output impedance all by itself.One thing I don't understand about his design, He has a bias voltage coming through P1. He could turn that up which would force him to increase the value of R4 to keep the same set current. That would be a good thing. R4 is not only the current set resistor but it is also acting as the "cathode" resistor if viewing the lower device (M2) as a cathode follower. The larger the value that resistor is, the more feedback there is for the device and that would lower the output impedance of the mu output.
The other thing I don't understand is this, "All schematics and designs are copyright Alejandro Moglia and Bartola Ltd. UK unless stated otherwise. All rights reserved. You are welcome to build the circuits presented here for your own personal entertainment. You may NOT build from information on this page for commercial profit without a royalty agreement with the author in place."
All a copyright prevents others from doing is coping the image of his schematic. Like making a copy of a copyrighted book and selling it. A copyright does not protect ideas. A circuit design is an idea. To protect a circuit design he would have to patent it. To patent it it would have to be new and unique. I'm not sure he's done anything that hasn't been done before.
Oh well, there are a lot of things I don't understand in this world.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 02/17/25
I worked for several years as a technical writer. You're correct regarding copyright. It's a violation to publish the same schematic in its original form (except under fair use provisions). However, if you redraw the schematic so that it's no longer the same document, all's well.
.
Have Fun and Enjoy the Music
"Still Working the Problem"
You can't drive into a 43 ohm load to ground with a stage that has a 8.3k output impedance. That would rotate the load line for the tube way to the vertical and there would be very little gain and a whole bunch of harmonic distortion.
8.3k output wants to see a 83k load.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Tre made the suggestion to use 68k+22k resistors to give me a dropping output level but not to limit upper frequency. Tried this ides with a 6N7 preamp I also had noise issues with and it worked well.Even with my DIY shieldless lead (175pF) into 100K upper 3db was 37KhZ. 6N7 are power tubes and Im not sure of they have a very high frequency range or not. Regardless I'm happy with 37K.
Applying the same 68k+20k idea to the 01a preamp gave me noise level outputs of 1.9mV and 0.9mV a big improvement over the ~4.0mV with the 300k+100k setup. At good volume only by putting my ear to my 93db sensitive 12" woofers do I hear the slightest hum then only in the left channel.
Good result for two preamps so far and two to go. Lucky I have a good collection of Rs to mod the others.
| retro-thermionic |
"Tried this ides with a 6N7 preamp I also had noise issues with and it worked well.Even with my DIY shieldless lead (175pF) into 100K upper 3db was 37KhZ. 6N7 are power tubes and Im not sure of they have a very high frequency range or not. Regardless I'm happy with 37K."
The 6N7 was primarily intended as a Class B output tube. But the 6N7 data sheets (GE, TungSol) suggests multiple uses for the tube including use as a Class A driver and gives suggested operating points for that application. Since nobody in the DIY community seems interested in building Class B amps all the forum discussion on the 6N7 seems to be about using it as an input / preamp tube.
Apparently, there are some glass (G, GT) versions. I have some of the 6A6s which are glass. The 6E6, 79 and 53 are supposedly equivalents or near equivalents. There's also the 5694 which is equivalent or nearly but has separate cathodes.
At one point I was considering building a Japanese SE 2A3 design that uses them on the input but I was going to use the 6A6 version. I like using less common tubes but I don't really have a need for a preamp with that much gain (mu 35).
I have priced on eBay some glass 6N7s and though not cheap well within my budget. They do look good though. But still having three boxes of tubes im trying not to buy more but use up what I have in new DIY gear and move on. I have a number of other metal tubes. All sound excellent in their respective amps.
| retro-thermionic |
Yeah, I don't remember if the 6A6s I got came from eBay or if I found them locally.
I'm less than 30 minutes from Radio Electric Supply (vacuumtubes.net) which is probably the biggest tube dealer in the world. They have warehouses full of tubes, many millions of them, though not all audio related, of course. They ship worldwide and have reasonable prices.
They're generally much cheaper than the "boutique" sellers who give flowery descriptions of how a particular tube allegedly sounds. As if a particular tube sounds the same regardless of which circuit it's used in or any other differences in the rest of the system. Of course, the "holy grail" tube$ are always what they just happen to have in $tock at any particular time.
The guys at RES (brothers, Roy and Dale) are not audiophiles, they just sell tubes. Hence the total lack of BS.
Another very large dealer is about 100 miles south of me in Orlando. That's vacuumtubesinc.com which also ships worldwide.
The only metal tube I've experimented with is the 6AG7. It caught my interest because it's incredibly easy to drive. I have a breadboarded PSE design that uses them with each channel driven by half of a 3A5, which is a dual DHT with a mu of 15.
It's another "inverted SET" (iSET) design which is a term coined by Andy Evans. I built another iSET amp which I call the Nuance. Details are in a thread on Audio Karma.
The "inverted" aspect refers to fact that instead of using DH output tubes driven either by higher mu indirectly heated tubes or multiple stages of somewhat lower mu tubes, the amp is only two stages with input tubes that are DH. The outputs are indirectly heated and to easy to drive, unlike the output tubes that are commonly used in SET amps.
The PSE 6AG7 is on the sidelines for now, though. I'm back to playing with low mu (10 or less) preamp tubes at the moment which is why your thread caught my eye.
"Tre made the suggestion to use 68k+22k resistors to give me a dropping output level but not to limit upper frequency."
Actually, I was the one who suggested using the combination of 68k in series and 22k to ground and said it should improve the -3db point of the upper frequencies.
My earlier post: "OK, so in the case of this 01A preamp, he could use a voltage divider with 68k in series and 22k to ground. This would reduce the voltage about the same as the 330k/100k combo he's using and it would present a load of 90k to the tube and be no problem.
The value of 68k || 22k is 16.62k. If we plug that into the low pass calculator then the -3db point would be 54.75k. That should also not be an issue, right?"
Tre quoted me and commented about the corresponding -1db point, which he said would be 27.375kHz. He suggested that, from a technical standpoint, this would still not be enough to "get the filter completely out of the audio band".
I'm glad to hear that, from a practical standpoint, it seems like it improved things subjectively. I try to understand the technical end of things - hence my questions in this thread - but I'm not convinced that striving for technical perfection is always audible beyond a certain point. Sometimes "better than it was" is good enough and going to more extreme lengths is just a waste of time and money if it doesn't yield audible improvement. IMO, but that's just me.
I guess for an engineer technical perfection IS the goal so it's a valid pursuit regardless of whether it's audible or not.
I understand how it's easy to get confused though. AA's software is ridiculously out of date and totally lame. Unlike other sites like DIY Audio, Audio Karma, and others AA doesn't allow you to easily quote a previous post without rewriting it and adding quotation marks. As I have done above.
Unless someone makes the effort and adds a quote manually it's often impossible to tell which previous post someone is responding to.
I suspect that AA will never modernize their software. I don't know about the site as a whole but the forums I used to visit regularly years ago are pretty much dead in the water now. Some of them now go weeks without a post where, in the past, there were a large number of new posts and threads every day. Sad.
I still find some useful and interesting posts on AA but I rarely post here anymore because of the lack of activity. Fortunately, the search function yields a wealth of information from the time when AA was a vibrant forum.
Well thanks for the suggestion. Probably like you Im on a number of forums mainly an Australian one which is where I live. Id like to think we are all here to listen and to help where we can fostering friendship and help in a common field of interest.
Two preamps to re-mod. Ill post results.
| retro-thermionic |
"You can't drive into a 43 ohm load to ground with a stage that has a 8.3k output impedance. . . . 8.3k output wants to see a 83k load."
OK, so in the case of this 01A preamp, he could use a voltage divider with 68k in series and 22k to ground. This would reduce the voltage about the same as the 330k/100k combo he's using and it would present a load of 90k to the tube and be no problem.
The value of 68k || 22k is 16.62k. If we plug that into the low pass calculator then the -3db point would be 54.75k. That should also not be an issue, right?
"OK, so in the case of this 01A preamp, he could use a voltage divider with 68k in series and 22k to ground. This would reduce the voltage about the same as the 330k/100k combo he's using and it would present a load of 90k to the tube and be no problem."Yes, you are getting this!
"The value of 68k || 22k is 16.62k. If we plug that into the low pass calculator then the -3db point would be 54.75k. That should also not be an issue, right?"
Well, almost. With a -3db point of 54.75kHz the -1db point will be 27.375kHz. That is out of the audio band but the phase is still being shifted by the filter. To get the filter completely out of the audio band the -3db point of the filter needs to be a full decade above 20kHZ.
***Also remember that these calculations do not take into account the Miller of the first stage of the power amp and that could be more capacitance than the cable.***
So to keep the filter totally out of the picture it needs a -3db point at 200kHz.
There are other things in a circuit that will not allow 200kHz bandwidth and there is not much we can do about it (200kHz would be almost impossible for an output transformer) but I don't see that as an excuse to not solve a problem that can be solved. These things stack up so there is no reason to exacerbate the problem.
My question would be, why does he need this gain reduction in the first place? Doesn't the preamp have a volume control in front of the tube stage?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 02/16/25 02/16/25
I like your 22k//68K idea. Yet to put the preamp on the CRO. Gain was reduced because there was a hum issue.
| retro-thermionic |
Are you using AC for the 5 volt filament?
You might want to try a critical inductance input choke filter DC supply for a DHT.
Or maybe try Rod Coleman's Filament Regulators.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Well filtered 6V then a 5V regulator. Also balanced 100ohm/100ohm network with Rk central tap. Hum is not in the filament voltage.
| retro-thermionic |
I guess not. So it is either the B+ or a hum field.What happens when you move the power supply chassis as far away as possible?
Do you have a schematic of the power supply?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 02/16/25 02/16/25
Finally got the preamp on the CRO. With the 300K/100K config, through the DIY shieldless cable (175pF), the upper 3db point was 19khZ into 100K. But, taking the 68K/22K suggestion lead to a much better upper 3db;- 37khZ. Not only that but the preamp was even quieter. One channel reading 0.0mV! Plaid all afternoon and throughly enjoyed it.Gain though is only 3db but these are power tubes so what do you expect? It will deliver 4Vpp into 100K. My DIY SE UL EL34 amp has plenty of gain so no concern. Not only that the two are a perfect synergy. Love the sound of this 6N7 preamp and love the metal tubes.
| retro-thermionic |
Edits: 02/26/25
"With the 300K/100K config, through the DIY shieldless cable (175pF), the upper 3db point was 19khZ into 100K. But, taking the 68K/22K suggestion lead to a much better upper 3db;- 37khZ. Not only that but the preamp was even quieter. One channel reading 0.0mV! Plaid all afternoon and throughly enjoyed it.
Gain though is only 3db but these are power tubes so what do you expect? It will deliver 4Vpp into 100K. My DIY SE UL EL34 amp has plenty of gain so no concern. Not only that the two are a perfect synergy. Love the sound of this 6N7 preamp and love the metal tubes."
6N7? Metal tubes? Huh?? Isn't this thread about a 01A preamp?
Looks like I got my threads mixed. But I've had the same problem over four different preamps based on the 01a, 71a, 6N7 and 12AX7 preamps. Forgive the mixup but the result (good result) hopefully will be the same.
| retro-thermionic |
It maybe an earth configuration issue. Ive used my design HT double Pi filtration in wellover 100 builds. Power amps and preamps. None have hums.
| retro-thermionic |
I see the preamp PS is earth grounded. Is the power amp earth grounded?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
"I like your 22k//68K idea."Yeah, if you need to place the gain reduction solution on the preamp I figured that would be a good compromise.
On the other hand, if you could put it on the amp end of the interconnect that would seem to be even better.
If I'm understanding this correctly, that would lower the output impedance of the preamp and result in a higher -3db rolloff point. Right??
Either a simple two resistor voltage divider or a volume pot could be mounted on the amp itself, at the input. This would be great if the preamp was always used with the same amp. But this seems like an unlikely scenario.
A more versatile solution would be to build the voltage divider into an interconnect which would be used exclusively with this preamp. Or any other pre that's either too noisy or has too much gain.
One question. If the voltage divider was put on the amp end of the interconnect would it be better to use much larger resistance values than 68k in series with 22k to ground? Wouldn't the 22k to ground be in parallel with the amp's existing input impedance (let's assume 100k)? Or would the 68k, 22k and 100k all be in parallel? Either way that would lower the input impedance significantly (to ~15k) and make the amp harder to drive.
It that's the case than you could use a combo of 2.2 meg in series and a 750k to ground to get about the same voltage/noise reduction. Or would this create other issues??
Obviously, if the hum is the result of a ground issue then you could eliminate the voltage divider as long as the increased gain is not an issue.
Edits: 02/17/25
All good ideas.
| retro-thermionic |
He could just fix the hum problem that started all of this and then he wouldn't need the attenuation at all. Impedance problem solved.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I tried for months but to no avail. This allows me to enjoy something I have spent a huge amount of time on.
| retro-thermionic |
I hate when that happens.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Sorry for asking more questions but I'm still confused and I'd like to understand how this calculation works. I do appreciate everyone's patience!
Here's a formula I was given a couple of years ago. It also considers the influence of a pot / voltage divider on the output. I mistakenly thought that the series resistance of the pot was only used to calculate the high pass value. Apparently, as was explained earlier, it's also a factor that's used to determine the output impedance.
(Tube's output impedance + series segment of pot or voltage divider) + (grounded segment of pot or voltage divider || input impedance of amp and/or other resistance to ground) = Overall output impedance
Tre's most recent calculation parallels the values used in the voltage divider, the 330k in series with the signal and the 100k that goes to ground and uses that to calculate the low pass frequency.
Why are they in parallel? According the formula above the 330k should be added to the impedance of the tube, not paralleled with the 100k that goes to ground.
Is the formula I was given incorrect or am I misinterpreting it somehow?
TK mentions that the tube's impedance should be added. The plate resistance of the 01A is 10k. As I understand it the tube's impedance is calculated by putting this in parallel with the tube's load resistor. If a 50k load resistor is used then the resulting impedance of the tube would be 8.33k.
Do we ignore the input impedance of the amp since it's on the other end of the cable? Or is it also part of the calculation?
The output impedance of a triode gain stage is the plate resistance in parallel with value of the plate resistor (the resistor between the power supply and the plate of the tube).
A triode needs to see a plate load resistor value of 3 to 5 times the tube plate resistance. (more if you have the supply voltage that will allow it or better yet a CCS)
A gain stage wants to see a load impedance 10 times the output impedance of the stage to prevent loading and rotating the load line for the triode towards the vertical causing a loss of gain and an increase of distortion.
When you drive a pot the total impedance of the pot is the load that is being driven by the preamp's output impedance (assuming there is no added resistance to ground following the wiper).
The output impedance of the pot (this is the value that is driving what ever follows the top) is dependent on the position of rotation.
If we assume that the impedance of that is driving the pot is zero (and of course it's not ), the output impedance of the pot at the -6db position is 1/4 of the value of the pot. So a 100k pot has an output impedance of 25k when the wiper is at the position with 50k above and 50k below it.
It would not take very much shunt capacitance to start disturbing the highs with a drive impedance of 25k.
So pots are a balancing act. Too low of a value and the source (the preamp in this example) has a hard time driving it, too high of a value and the pot has a had time driving what it is trying to drive.
All of the numbers that describe all of this are easily found on the internet.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
This post is made possible by the generous support of people like you and our sponsors: