Bas, you beat me to this.
So far it seems to be a positive sonically (i haven't heard it, but Chris did it to his Jolida PP EL34, and Geek is trying it in his/my DynaMutt). Daniel has concerns about the undefined DC operating point and what happens when things go sideways.
I posted a similar inquiry on the JoeList and there has been one interesting reply. A similar concern with DC op point, but also some speculation about what is happening.
A high value R from suppressor to ground (or to a negative bias*) could be the fix for the DC op point.
The suppressor to negative is a technique that has been touted by Bill Perkins for some time, and Eddie Vaughn pointed out that this was used with good results in a Traynor "copy" of a Marshall from the 70s.
dave
So far it seems to be a positive sonically (i haven't heard it, but Chris did it to his Jolida PP EL34, and Geek is trying it in his/my DynaMutt). Daniel has concerns about the undefined DC operating point and what happens when things go sideways.
I posted a similar inquiry on the JoeList and there has been one interesting reply. A similar concern with DC op point, but also some speculation about what is happening.
A high value R from suppressor to ground (or to a negative bias*) could be the fix for the DC op point.
The suppressor to negative is a technique that has been touted by Bill Perkins for some time, and Eddie Vaughn pointed out that this was used with good results in a Traynor "copy" of a Marshall from the 70s.
dave
Here is Phil's JoeList post
I also have some SV83 (6P15P) on order so we can try this on our El Cheapo variant.
dave
I also have some SV83 (6P15P) on order so we can try this on our El Cheapo variant.
dave
Being the engineer I am, I have to wonder if this "sonic improvement" they are hearing is simply an operating point shift. I would like to see the G3 voltage and plate current, before and after...
Pete
Pete
Is it really shifting the operating point?
Does the suppressor grid really have much effect on the DC op point?
The cap is in series with the suppressor grid, which is normally tied to the cathode. This takes the suppressor grid out of the circuit at DC and low frequencies depending on the cap value (which they don't specify).
It seems like it's function (the suppressor and this mod) is a large signal AC phenomenon. The cap tunes in when the suppressor takes effect, in frequency.
You are in effect tuning the maximum energy transfer in the frequency domain.
Kind of kinky
Does the suppressor grid really have much effect on the DC op point?
The cap is in series with the suppressor grid, which is normally tied to the cathode. This takes the suppressor grid out of the circuit at DC and low frequencies depending on the cap value (which they don't specify).
It seems like it's function (the suppressor and this mod) is a large signal AC phenomenon. The cap tunes in when the suppressor takes effect, in frequency.
You are in effect tuning the maximum energy transfer in the frequency domain.
Kind of kinky
W...T...F?
This is the audio equivalent of foot binding. It has no practical effect, may be detrimental, and serves only to entertain those silly enough to follow the fad.
This is the audio equivalent of foot binding. It has no practical effect, may be detrimental, and serves only to entertain those silly enough to follow the fad.
Varying g3 voltage on a dual control pentode certainly has some effect. These have a more dense g3 winding than normal pentodes, so the effect should be larger there. Positive voltage on g3 for a 6LE8, for example, squares up the characteristic curves (+12V to +15V on g3 giving very square Mosfet like curves, but go too positive and the g3 sucks up current itself). A negative voltage rounds the corner off the curves.
http://scottbecker.net/tube/sheets/135/6/6LE8.pdf
A positve g3 also lowers screen current since more electrons make it to the plate. A negative g3 will raise it, maybe dangerously. I would suggest not only measuring the g3 voltage with the g3 cap, but also check g2 current.
Don
http://scottbecker.net/tube/sheets/135/6/6LE8.pdf
A positve g3 also lowers screen current since more electrons make it to the plate. A negative g3 will raise it, maybe dangerously. I would suggest not only measuring the g3 voltage with the g3 cap, but also check g2 current.
Don
Last edited:
Being the audiophile I am. I wonder if this "sonic improvement" is nothing more than a marketing statement. BUT I'm always willing to try stuff and see for myself. Which I will try some time soon. And will report back.
In my measurements on Chris' Jolida 302, I found as was expected that the space charge was pulling the g3 down. Unless of course some other phenomenon is doing it.
I would surmise as others have that simply moving g3 around but with definite control over its DC point should have an effect on the 'sound' of the tube by effectively setting it into a different operating point. I can't say that a capacitor would necessarily be any better than just biasing with a traditional supply of some sort.
I do however wonder how the capacitor affects things during/after clipping. Let's say the output stage plate(s) has been driven heavily negative, say until there's g1 current... Perhaps this condition leaves the capacitor charged to a different state than when at idle and causes an interesting recovery that is supposed to sound better during high output? Who knows, this is only speculation.
Regardless of all else; my main concern was the indefinite bias for g3.
I make no judgment on the sonic benefit as I have no informed reason to believe that it should improve or degrade anything and have not listened to music through amplifiers modified in this way enough to determine any noticeable difference.
I would surmise as others have that simply moving g3 around but with definite control over its DC point should have an effect on the 'sound' of the tube by effectively setting it into a different operating point. I can't say that a capacitor would necessarily be any better than just biasing with a traditional supply of some sort.
I do however wonder how the capacitor affects things during/after clipping. Let's say the output stage plate(s) has been driven heavily negative, say until there's g1 current... Perhaps this condition leaves the capacitor charged to a different state than when at idle and causes an interesting recovery that is supposed to sound better during high output? Who knows, this is only speculation.
Regardless of all else; my main concern was the indefinite bias for g3.
I make no judgment on the sonic benefit as I have no informed reason to believe that it should improve or degrade anything and have not listened to music through amplifiers modified in this way enough to determine any noticeable difference.
One of the main advantages that solid state has is the much improved gain without having all those extra grids to deal with. Screen and suppressor grids are also auxilliary control grids, and what you hang on them can make a big difference in circuit characteristics. Usually, your best bet is to get as much impedance out of your screen circuits as possible.
However, this article looks like typical audiophoolery: "We did this thing, and it's the mostest bestest mod evar!" No before and after data, no details on what this does to the suppressor voltage (seems that it wouldn't be any different from just letting it float -- at least so far as DC is concerned) nothing concerning how this affects bias points, or plate current.
What I do know is that negative suppressor -- cathode voltages are definitely not good, unless it's your intention to make a UHF, dynatron-type oscillator. That will cause a negative resistance characteristic to appear at the screen grid. At the very least, it looks like an open invitation to the instability gremlins to take up permanent residence there. Positive voltages on suppressors/beam formers has long been the cure for "snivets", yet another phenomenon caused by the appearance of negative resistances in your screen circuit.
However, this article looks like typical audiophoolery: "We did this thing, and it's the mostest bestest mod evar!" No before and after data, no details on what this does to the suppressor voltage (seems that it wouldn't be any different from just letting it float -- at least so far as DC is concerned) nothing concerning how this affects bias points, or plate current.
What I do know is that negative suppressor -- cathode voltages are definitely not good, unless it's your intention to make a UHF, dynatron-type oscillator. That will cause a negative resistance characteristic to appear at the screen grid. At the very least, it looks like an open invitation to the instability gremlins to take up permanent residence there. Positive voltages on suppressors/beam formers has long been the cure for "snivets", yet another phenomenon caused by the appearance of negative resistances in your screen circuit.
This is the decoy: g3 is often described as another grid, but the same misnomer is used as beam confining plates with virtually nil effect on the electron stream. The data sheets don't always mention this.
The 6CH6,12BY7 breed have confining plates.
richy
Daniel, did you take any further operating point measurements when you repaired my minor wiring mistake the other day?
If not, would you like to poke around again?
That data might answer some of the questions listed above, but of course not necessarily :
"what exactly does it do for the sonics?"
As I read it, some of the reputed benefits are "opening up" of soundstage, dimensionality, etc. - and unless it'd be safe to install a switch to short out the caps and return to the standard strapped suppressor grid configuration, a quick test for the difference with a single amp would be kinda tricky. To be honest the Jolida if by far the least listened to amp in my collection, it's been months since it's even been turned on, so even "sonic memories" have long since faded
"what could possibly go wrong?"
is this another potential answer to the multiple choice list of "how many ways are there to release expensive smoke from this unit?"
"why would a small manufacturer like Deckert include this in all new production amps using the SV83?" He's known for pretty good customer service, and I'd assume has considered the warranty implications
If not, would you like to poke around again?
That data might answer some of the questions listed above, but of course not necessarily :
"what exactly does it do for the sonics?"
As I read it, some of the reputed benefits are "opening up" of soundstage, dimensionality, etc. - and unless it'd be safe to install a switch to short out the caps and return to the standard strapped suppressor grid configuration, a quick test for the difference with a single amp would be kinda tricky. To be honest the Jolida if by far the least listened to amp in my collection, it's been months since it's even been turned on, so even "sonic memories" have long since faded
"what could possibly go wrong?"
is this another potential answer to the multiple choice list of "how many ways are there to release expensive smoke from this unit?"
"why would a small manufacturer like Deckert include this in all new production amps using the SV83?" He's known for pretty good customer service, and I'd assume has considered the warranty implications
Last edited:
I didn't take any more measurements than to see what voltage the suppressors ended up at. This was of course loaded by the 10Mohm of my DMM and of course might have been a much larger value when unloaded. The voltage was about -600mV g3-k.
I could take the amplifier back and make some more measurements just to see if I find anything. It's up to you and anyone else who would like to find some data.
I could take the amplifier back and make some more measurements just to see if I find anything. It's up to you and anyone else who would like to find some data.
Steve has a followup article on the Hazen mod with some measurements here:
DECWARE - The Hazen Mod
Dean
DECWARE - The Hazen Mod
Dean
I have not tried this so I will not comment on its merit. I will however offer this observation based on my experiences with current production and NOS tubes.
Leaving any grid without a low impedance path to a desired potential may be an invitation to disaster on many current production tubes. The tubes that were produced 20+ years ago generally had a much harder vacuum than the stuff being produced today. Many of todays power tubes routinely show signs of ionized gas inside the plate area. This can be measured as a small grid current. Find a cathode biased amp with a high valued grid (G1) resistor connect a voltmeter to G1 and then measure the voltage on G1 with several tubes. It should be slightly negative due to electron bombardment. It may show zero or slightly positive on some new tubes, and this will change with operating temperature. What voltage could you expect on a grid with no DC path to ground? I don't know, but I bet that it won't be the same under all conditions with all tubes!
As Miles stated a negative voltage on G3 can produce oscillation, and the reason G3 is pinned out seperately on many sweep tubes is so that a fixed positive voltage could be applied to cure barkhausen oscillation or "snivets". This is clearly mentioned on the data sheets for many sweep tubes.
There was however another thread here about a year ago where the poster claimed that a small negative voltage applied to G3 on an EL34 worked magic for the sound. It made the grid "dissapear". I never got around to trying it, although some of his other claims had some technical merit.
Leaving any grid without a low impedance path to a desired potential may be an invitation to disaster on many current production tubes. The tubes that were produced 20+ years ago generally had a much harder vacuum than the stuff being produced today. Many of todays power tubes routinely show signs of ionized gas inside the plate area. This can be measured as a small grid current. Find a cathode biased amp with a high valued grid (G1) resistor connect a voltmeter to G1 and then measure the voltage on G1 with several tubes. It should be slightly negative due to electron bombardment. It may show zero or slightly positive on some new tubes, and this will change with operating temperature. What voltage could you expect on a grid with no DC path to ground? I don't know, but I bet that it won't be the same under all conditions with all tubes!
As Miles stated a negative voltage on G3 can produce oscillation, and the reason G3 is pinned out seperately on many sweep tubes is so that a fixed positive voltage could be applied to cure barkhausen oscillation or "snivets". This is clearly mentioned on the data sheets for many sweep tubes.
There was however another thread here about a year ago where the poster claimed that a small negative voltage applied to G3 on an EL34 worked magic for the sound. It made the grid "dissapear". I never got around to trying it, although some of his other claims had some technical merit.
- Status
- Not open for further replies.