VHT D-50 Mod Journey
Moderators: pompeiisneaks, Colossal
-
Reeltarded
- Posts: 9955
- Joined: Sat Feb 14, 2009 4:38 am
- Location: GA USA
1 others liked this
Re: VHT D-50 Mod Journey
Waste of a switch.
Signatures have a 255 character limit that I could abuse, but I am not Cecil B. DeMille.
Re: VHT D-50 Mod Journey
Aha nice! So basically you felt like removing local nfb on v1b was an improvement? How did it impact the amp? Breaks up a little easier when clean? Did it sweeten the breakup or change the character?
My next mod is to put a 100k log pot as variable resistor in series with the feedback resistor on power section - hoping that makes it feel less sterile even if not breakup up amp - we will see
My next mod is to put a 100k log pot as variable resistor in series with the feedback resistor on power section - hoping that makes it feel less sterile even if not breakup up amp - we will see
-
Reeltarded
- Posts: 9955
- Joined: Sat Feb 14, 2009 4:38 am
- Location: GA USA
Re: VHT D-50 Mod Journey
All that is done so far is cut LNFB out and EH preamp tubes. The tube change made me sure that these things are hyper sensitive to each tube position as they should be. The swamped bass is gone, the smeared muffling is cleared. I would say I can actually identify the effects of the signal caps by position, audibly. It's got a really nice sing with the clean @4-5 and the OD section at similar. That seems closely balanced in the output as well.
Ditched LNFB added 30% body and as much gain. Warmth immediately had to be moderated by dropping the bass a number or so and adding treble.
NOTE: the general opinion was the amp turned into a VOX (by comparison) when we swapped preamps tubes it came so far mid-forward. (good thing)
The peaker after the first stage needs to be a stiff mica or a dense disk. The amp shows virtual indifference to your tone setting choices. I don't hate the snubbers on the OD section. You can't make a painful choice of treble. It's smooth.
The tone stack lacks the authority I am used to. Meh.
I am going to drop the output grids and change the resistor values in and out f the OD section then balance the new PI tube.The amp gained sustain in one way and lost a little in another.
In my opinion, these things sound like absolute shit out of the box. This can only mean 4-500 dollar ones in the future. The expectation is NOT MET. $50 and a set of end cutters later you have one of the best new amps I have played in forever.
That PAB.. oof. The external entrance trim was useful during the tube change.
Ruby tubes are awful. EH tubes are kind and decent.
Ditched LNFB added 30% body and as much gain. Warmth immediately had to be moderated by dropping the bass a number or so and adding treble.
NOTE: the general opinion was the amp turned into a VOX (by comparison) when we swapped preamps tubes it came so far mid-forward. (good thing)
The peaker after the first stage needs to be a stiff mica or a dense disk. The amp shows virtual indifference to your tone setting choices. I don't hate the snubbers on the OD section. You can't make a painful choice of treble. It's smooth.
The tone stack lacks the authority I am used to. Meh.
I am going to drop the output grids and change the resistor values in and out f the OD section then balance the new PI tube.The amp gained sustain in one way and lost a little in another.
In my opinion, these things sound like absolute shit out of the box. This can only mean 4-500 dollar ones in the future. The expectation is NOT MET. $50 and a set of end cutters later you have one of the best new amps I have played in forever.
That PAB.. oof. The external entrance trim was useful during the tube change.
Ruby tubes are awful. EH tubes are kind and decent.
Signatures have a 255 character limit that I could abuse, but I am not Cecil B. DeMille.
Re: VHT D-50 Mod Journey
Your experience dropping the local nfb makes sense with what modeling says - the cap / resistor parallel combo on grid of v1b is actually a treble shelf (cut), but then add the feedback and because resistance is higher in bass frequencies across those components you get more feedback in the bass, so it is flatish with a bit of treble cut but more feedback in bass than treble.
I need to post some more stuff, but I did measure the amps distortion last night and with all feedback in place it is around 1%, which is really low and is likely going to be swamped by speaker distortion, so it is almost solid statey in linearity and low distortion with stock amount of feedback run clean. Obviously tone stack makes frequency response very different. Anyhow surprisingly linear.
Other note - when swapping power tubes (I also put in EH - why not) I noticed the amp was biased pretty cold as stock. Worth checking power tube bias but I'm sure you did that. I will say that JJ stock tube looked pretty beefy inside - I have some mesa 6l6s in hand and the EH and all 3 were clearly physically different, which is nice as some have claimed all modern tubes are the same - hey if they are built differently they cannot be exactly the same!
I need to post some more stuff, but I did measure the amps distortion last night and with all feedback in place it is around 1%, which is really low and is likely going to be swamped by speaker distortion, so it is almost solid statey in linearity and low distortion with stock amount of feedback run clean. Obviously tone stack makes frequency response very different. Anyhow surprisingly linear.
Other note - when swapping power tubes (I also put in EH - why not) I noticed the amp was biased pretty cold as stock. Worth checking power tube bias but I'm sure you did that. I will say that JJ stock tube looked pretty beefy inside - I have some mesa 6l6s in hand and the EH and all 3 were clearly physically different, which is nice as some have claimed all modern tubes are the same - hey if they are built differently they cannot be exactly the same!
-
Reeltarded
- Posts: 9955
- Joined: Sat Feb 14, 2009 4:38 am
- Location: GA USA
Re: VHT D-50 Mod Journey
Oh, when in Jazz mode, very unexpected. Gain drops about half and a humbucker sounds tiny and terrible.
Yes. That LNFB is the wrong amplitude at the wrong frequencies.
Your statement on the power section lends itself very well to my idea than a Dumble is a great candidate for FET power section. Pretty sure it doesn't matter.. generally.
Yes. That LNFB is the wrong amplitude at the wrong frequencies.
Your statement on the power section lends itself very well to my idea than a Dumble is a great candidate for FET power section. Pretty sure it doesn't matter.. generally.
Signatures have a 255 character limit that I could abuse, but I am not Cecil B. DeMille.
Re: VHT D-50 Mod Journey
Yeah it is hard for me to see ever using the jazz setting. I'm sure I could change around the tone controls to get a usable setting but...why? I'll measure it all and we can look at it, but not seeing jazz used in my future and want that switch real estate to be used for something else useful (like a bass cut going into OD section, which I do want).
Looking at the data sheets for 6L6 vs say EL34 even without feedback the 6L6 datasheet indicates 1-2 percent distortion as being normal for class AB - EL34 showing more like 9 percent - which makes sense as you get some compression before distortion with the 34 given the rounder knee, whereas the 6L6 is a bit more linear until it hits the wall at low voltage. So yes essentially if you went Dumble preamp to solid state power section vs stock D power section I don't think you would notice much difference, absent the presence control which changes the frequency response and (in theory) will loosen up the top a bit
Looking at the data sheets for 6L6 vs say EL34 even without feedback the 6L6 datasheet indicates 1-2 percent distortion as being normal for class AB - EL34 showing more like 9 percent - which makes sense as you get some compression before distortion with the 34 given the rounder knee, whereas the 6L6 is a bit more linear until it hits the wall at low voltage. So yes essentially if you went Dumble preamp to solid state power section vs stock D power section I don't think you would notice much difference, absent the presence control which changes the frequency response and (in theory) will loosen up the top a bit
Re: VHT D-50 Mod Journey
Another benefit of cutting the feedback - those two huge resistors (40 meg total going by memory) have to be adding some noise - will see if I can pick up the difference if I cut it
Re: VHT D-50 Mod Journey
I don't think I've posted this yet - response curves of the grid stopper / cap combo with feedback:
Ok a bit of explanation - this is my modded version with the tone stack later, made into four identical copies in ltspice and with some tweaks made to each. In every one the circuit response is probed where the red "x" is (after V1B coupling cap). Bright cap out of circuit, gain on half (900/100K)
Green: The stock circuit with feedback and the 220K / 500pF "grid stopper" combo
Purple: Same as green, but cut the feedback out
Red: Take purple and now delete the 500pF cap across the 220K grid stopper
Teal: Take Red and delete the grid stopper (so just a normal gain stage with no feedback or grid stopper)
Commentary:
- The green vs purple shows the magnitude of the feedback - mostly in the bass frequencies - clearly shows the high cut shelf that is the 220K/500pf combo without the feedback.
- The purple vs red shows how much high would have been lost without the 500pF parallel cap - guessing the amp would be pretty dull
- Teal vs green shows how much gain total is available if wanted - about 3db across the board, with a bit less on the high end
To me the most promising curves for a clean channel are green and teal. Unless I was always playing really bright guitars I think the purple shelfed version which is just the feedback delete would be too much. That said as the amp has a smiley face eq perhaps this would manifest as a mid boost when you roll off the bass as reel did, or maybe you would just always use a bright switch - will have to measure
Ok a bit of explanation - this is my modded version with the tone stack later, made into four identical copies in ltspice and with some tweaks made to each. In every one the circuit response is probed where the red "x" is (after V1B coupling cap). Bright cap out of circuit, gain on half (900/100K)
Green: The stock circuit with feedback and the 220K / 500pF "grid stopper" combo
Purple: Same as green, but cut the feedback out
Red: Take purple and now delete the 500pF cap across the 220K grid stopper
Teal: Take Red and delete the grid stopper (so just a normal gain stage with no feedback or grid stopper)
Commentary:
- The green vs purple shows the magnitude of the feedback - mostly in the bass frequencies - clearly shows the high cut shelf that is the 220K/500pf combo without the feedback.
- The purple vs red shows how much high would have been lost without the 500pF parallel cap - guessing the amp would be pretty dull
- Teal vs green shows how much gain total is available if wanted - about 3db across the board, with a bit less on the high end
To me the most promising curves for a clean channel are green and teal. Unless I was always playing really bright guitars I think the purple shelfed version which is just the feedback delete would be too much. That said as the amp has a smiley face eq perhaps this would manifest as a mid boost when you roll off the bass as reel did, or maybe you would just always use a bright switch - will have to measure
You do not have the required permissions to view the files attached to this post.
Re: VHT D-50 Mod Journey
One thing I'll put here, but really it should probably be its own thread, is that I did some modeling of LED bias and how it would compare to bypassing a resistor with a large capacitor. The reason I modeled it was, at first blush, it seemed like a large bypass cap with resistor and LED (of same bias voltage) would be the same - and frequency response wise, they are the same, but I figured there would be a difference in the cathode voltage over time, and sure enough, when you model it in spice, it shows up.
Basic setup was three identical (other than cathode) preamp sections all set to bias at ~2.2V and the gain turned up to clip V1B. I tried a bunch of variants but the main three explored here are:
- Diode bias (2 diodes in series to get the right bias, a MURS120 and a QTLP690C LED. I don't know much about LEDs - these were chosen because they almost matched the voltage of the stock V1B in the D50 and there was an existing ltspice model for them
- 2.55K parallel with 4.7u (fully bypassed)
- 2.55K parallel with 47u (reaaaaaly bypassed)
Basic circuit used (red x is on cathode we are messing with), which is the same as my modded (later tone stack) D50 but with 2.55K resistor instead of 2.5K Learnings:
- LED is NOT a perfect cap at certain voltage (measuring voltage right above diode) – can drop below and go a bit above - see image below
--- Didn’t expect that
--- Not sure how accurate spice model is but its probably decent
- LED has slightly less gain than fully bypassed stage
--- Likely due to some effective series resistance in LED? I cannot explain this
--- Not much though - <1db
--- Big difference is that fully bypassed stages cathode voltage goes up as play sustained note
------ Effectively increases positive side headroom and moves bias position
------ The smaller the cap, the faster it goes to this position (assuming it still fully bypasses those frequencies; if too small, it doesn’t have DC excursion – sine wave at cathode just looks like unbypassed case, which doeesn't budge off of DC bias)
------ Bigger cap takes longer and more effectively bypasses everything (e.g. even bypassed there is some feedback at lower frequencies, even if not much)
The below shows the voltage at the cathode of V1B at normal (I think called quiescent) bias and then when a 240hz sine wave is played for 0.2s. On the chart blue is LED, red is 2.55K parallel 4.7u, green is 2.55K parallel 47u. As you can see above, the 4.7u bypass cap gets to its higher cathode voltage quickly - within a few cycles - and has a decent amount of ripple as it is bypassed but not perfectly bypassed (I think - open to other interpretations here). The green 47u takes a while to rise, but then also a while to fall after the note is played (which is the cause of the infamous blocking distortion when much more distorted). The blue is the diode. I thought the blue line would be just a perfect straight line - what would cause the diode to behave like this? I can see how you could get slightly above 2.2v if there is some effective series resistance, but what would cause it to fully or partially "turn off"? Too little current? I'm very curious if anyone has any answers on this.
In any case - a large cap bypassed resistor is not the same as a really large bypass cap is not the same as a diode for bias, at least in the time domain. I would imagine the shift in bias point would have some kind of dynamic feel and playing effect for the bypassed ones - and while I always used to think "what is the smallest cap I can get away with" to limit blocking distortion, now I'm going to start thinking about (like a compressor) what attack and release "times" give me what I think I want in the circuit. Anyhow - very interested in thoughts and reactions on this one
Basic setup was three identical (other than cathode) preamp sections all set to bias at ~2.2V and the gain turned up to clip V1B. I tried a bunch of variants but the main three explored here are:
- Diode bias (2 diodes in series to get the right bias, a MURS120 and a QTLP690C LED. I don't know much about LEDs - these were chosen because they almost matched the voltage of the stock V1B in the D50 and there was an existing ltspice model for them
- 2.55K parallel with 4.7u (fully bypassed)
- 2.55K parallel with 47u (reaaaaaly bypassed)
Basic circuit used (red x is on cathode we are messing with), which is the same as my modded (later tone stack) D50 but with 2.55K resistor instead of 2.5K Learnings:
- LED is NOT a perfect cap at certain voltage (measuring voltage right above diode) – can drop below and go a bit above - see image below
--- Didn’t expect that
--- Not sure how accurate spice model is but its probably decent
- LED has slightly less gain than fully bypassed stage
--- Likely due to some effective series resistance in LED? I cannot explain this
--- Not much though - <1db
--- Big difference is that fully bypassed stages cathode voltage goes up as play sustained note
------ Effectively increases positive side headroom and moves bias position
------ The smaller the cap, the faster it goes to this position (assuming it still fully bypasses those frequencies; if too small, it doesn’t have DC excursion – sine wave at cathode just looks like unbypassed case, which doeesn't budge off of DC bias)
------ Bigger cap takes longer and more effectively bypasses everything (e.g. even bypassed there is some feedback at lower frequencies, even if not much)
The below shows the voltage at the cathode of V1B at normal (I think called quiescent) bias and then when a 240hz sine wave is played for 0.2s. On the chart blue is LED, red is 2.55K parallel 4.7u, green is 2.55K parallel 47u. As you can see above, the 4.7u bypass cap gets to its higher cathode voltage quickly - within a few cycles - and has a decent amount of ripple as it is bypassed but not perfectly bypassed (I think - open to other interpretations here). The green 47u takes a while to rise, but then also a while to fall after the note is played (which is the cause of the infamous blocking distortion when much more distorted). The blue is the diode. I thought the blue line would be just a perfect straight line - what would cause the diode to behave like this? I can see how you could get slightly above 2.2v if there is some effective series resistance, but what would cause it to fully or partially "turn off"? Too little current? I'm very curious if anyone has any answers on this.
In any case - a large cap bypassed resistor is not the same as a really large bypass cap is not the same as a diode for bias, at least in the time domain. I would imagine the shift in bias point would have some kind of dynamic feel and playing effect for the bypassed ones - and while I always used to think "what is the smallest cap I can get away with" to limit blocking distortion, now I'm going to start thinking about (like a compressor) what attack and release "times" give me what I think I want in the circuit. Anyhow - very interested in thoughts and reactions on this one
You do not have the required permissions to view the files attached to this post.
Re: VHT D-50 Mod Journey
I mentioned before I took some direct measurements of frequency response and distortion of the amp - here's what I did and the results.
Setup:
- Focusrite Scarlett 18i20 interface used for I/O
- Output of scarlett into input of amp. Amp into custom load box (has circuit inside which is the electrical equivalent of a speaker, so resonates, etc., but electrically)
- Load box into scarlett
Ok let's first do a loop on the scarlett to see frequency response and distortion:
Ok so based on the ruler flat response and very low distortion - the interface won't be contributing to the measurements at all.
All of the below unless mentioned otherwise are with the clean channel, lower signal input (no clipping), everything at "noon", rock setting, no boosts, bright switch off.
Here is the effect of the bright switch: Up starts boosting earlier and has more boost. Down is more subtle and only really kicks in at 2-3K
Here is the effect of the mid switch and PAB switches: The mid switch is really nice - blows away the dip but doesn't really move the bass/treble. Would be nice to have that on a footswitch - would really change the amp's character. Coming from playing a lot of tweed style amps with the one knob tone control I like the sound of the mid lifted (similar frequency response). The PAB is really a different animal - reminds me of the shape of the Klon Centaur response curve (see electrosmash analysis here: https://www.electrosmash.com/klon-centaur-analysis) and adds a lot of gain. Putting them both together you get yet more gain. So...I still stand by putting a tone stack before the power amp...but...seeing these curves I think I am missing out. With this kind of mid shaping pre-distortion you can really hit that distortion section hard and shape the tone. As is this pre-amp boost, which would be great into distortion, is kind of wasted where I put it right before the power amp (unless you are trying to clip your power amp - that is another option/tone).
Here is the range of the treble control: Kind of overlappy with the presence. Really presence is like "high treble" - like having a 4th tone knob that only adds. I might consider cutting more in the presence band just so I can selectively add it back as desired (more range/options).
Here is the range of the mid control: Kind of weak IMO.
Here is the range of the bass control: Really usable - especially the cut. Again if running dirty only and cranking the gain you would probably have to cut bass way back (or hit that PAB to cut it for you).
Here is the range of the presence control: I feel vindicated here! I felt crazy like the presence control didn't do anything until noon, and sure enough, it really doesn't do anything until after noon. I can rest easy now
Here is rock vs jazz - jazz cuts 10-15db off an already lossy tone stack - no thanks. I'll play with jazz settings later to see if there are interesting curves hiding in there but I think that switch is getting repurposed.
I also added a 100k log pot in series with the feedback resistor - sized it so that at min it would be "stock" and at max it would be essentially no feedback. Here are the results at different knob locations: It works as intended. Worth noting that there is less feedback in the midrange and more in the bass/treble region. I believe this is because the reactive load has a higher impedance in treble and bass (as with a real speaker) so the open loop gain of the power section is higher there, and so the amount of feedback is higher. Anyhow regardless it is interesting.
Setup:
- Focusrite Scarlett 18i20 interface used for I/O
- Output of scarlett into input of amp. Amp into custom load box (has circuit inside which is the electrical equivalent of a speaker, so resonates, etc., but electrically)
- Load box into scarlett
Ok let's first do a loop on the scarlett to see frequency response and distortion:
Ok so based on the ruler flat response and very low distortion - the interface won't be contributing to the measurements at all.
All of the below unless mentioned otherwise are with the clean channel, lower signal input (no clipping), everything at "noon", rock setting, no boosts, bright switch off.
Here is the effect of the bright switch: Up starts boosting earlier and has more boost. Down is more subtle and only really kicks in at 2-3K
Here is the effect of the mid switch and PAB switches: The mid switch is really nice - blows away the dip but doesn't really move the bass/treble. Would be nice to have that on a footswitch - would really change the amp's character. Coming from playing a lot of tweed style amps with the one knob tone control I like the sound of the mid lifted (similar frequency response). The PAB is really a different animal - reminds me of the shape of the Klon Centaur response curve (see electrosmash analysis here: https://www.electrosmash.com/klon-centaur-analysis) and adds a lot of gain. Putting them both together you get yet more gain. So...I still stand by putting a tone stack before the power amp...but...seeing these curves I think I am missing out. With this kind of mid shaping pre-distortion you can really hit that distortion section hard and shape the tone. As is this pre-amp boost, which would be great into distortion, is kind of wasted where I put it right before the power amp (unless you are trying to clip your power amp - that is another option/tone).
Here is the range of the treble control: Kind of overlappy with the presence. Really presence is like "high treble" - like having a 4th tone knob that only adds. I might consider cutting more in the presence band just so I can selectively add it back as desired (more range/options).
Here is the range of the mid control: Kind of weak IMO.
Here is the range of the bass control: Really usable - especially the cut. Again if running dirty only and cranking the gain you would probably have to cut bass way back (or hit that PAB to cut it for you).
Here is the range of the presence control: I feel vindicated here! I felt crazy like the presence control didn't do anything until noon, and sure enough, it really doesn't do anything until after noon. I can rest easy now
Here is rock vs jazz - jazz cuts 10-15db off an already lossy tone stack - no thanks. I'll play with jazz settings later to see if there are interesting curves hiding in there but I think that switch is getting repurposed.
I also added a 100k log pot in series with the feedback resistor - sized it so that at min it would be "stock" and at max it would be essentially no feedback. Here are the results at different knob locations: It works as intended. Worth noting that there is less feedback in the midrange and more in the bass/treble region. I believe this is because the reactive load has a higher impedance in treble and bass (as with a real speaker) so the open loop gain of the power section is higher there, and so the amount of feedback is higher. Anyhow regardless it is interesting.
You do not have the required permissions to view the files attached to this post.
Re: VHT D-50 Mod Journey
I am the same way, it's not till around 1:00 that I hear anything move as well. Thanks for the plots!I feel vindicated here! I felt crazy like the presence control didn't do anything until noon, and sure enough, it really doesn't do anything until after noon. I can rest easy now
Tony
" The psychics on my bench is the same as Dumble'"
Re: VHT D-50 Mod Journey
unfortunatelly, again, all those simulations are not helping too much with the real amp. if the amp has dry or farting / fuzzy overdrive sound, no any sim will show this ot you know what you need to do. or the worst case, amp works but mediocre in all ways. this is due to everything, tubes, transformers , lead dress and components used. no any fast electrical intervention will help here. i'm playing last 3 weeks with 2 Dumble builds 10 years old, both were useless. One i get to sound fantastic, with changing components what would be Dumble selection, transformers and tubes were ok from before, but with other, after too much tweaking, i cant move too much and will be dissasembled, rebuilt.
-
Reeltarded
- Posts: 9955
- Joined: Sat Feb 14, 2009 4:38 am
- Location: GA USA
2 others liked this
Re: VHT D-50 Mod Journey
While waiting for back orders, I started pushing wires around at V1/V2.
It seems insane that small angle changes do so much. Stunning. The wires should be on knobs. lmao
At medium gain settings the OD sings like a decently trained bird even though the overall sound is more than a little congested. Wow.
It seems insane that small angle changes do so much. Stunning. The wires should be on knobs. lmao
At medium gain settings the OD sings like a decently trained bird even though the overall sound is more than a little congested. Wow.
Signatures have a 255 character limit that I could abuse, but I am not Cecil B. DeMille.
-
Reeltarded
- Posts: 9955
- Joined: Sat Feb 14, 2009 4:38 am
- Location: GA USA
1 others liked this
Re: VHT D-50 Mod Journey
Operation Dumbol Stop is about to proceed. All inmplements and affects in-hand.
I am going to start by replacing anything in the signal path that is non-Dumble value. The next things are cathodes and plates. The next step is cap then resistors.
Woo. I'm excited to hear the face of this thing.
I am going to start by replacing anything in the signal path that is non-Dumble value. The next things are cathodes and plates. The next step is cap then resistors.
Woo. I'm excited to hear the face of this thing.
Signatures have a 255 character limit that I could abuse, but I am not Cecil B. DeMille.
Re: VHT D-50 Mod Journey
Nice! I'm looking forward to hearing how it comes out. I put my v1b local negative feedback on a switch today (sacrificed the manual preamp boost switch). There is a real difference in the amp though a non-guitarist probably wouldn't hear it (it is a bit louder but in talking tone). Best way I can out it - slightly more warm and hairier - slightly smoother transition into clean tone overdrive. If you want the spanky clean sound the with feedback setting is better - but I tend to like more warmth and hair so no feedback is my jam. Fyi I put in a 50k grid stopper when no feedback is engaged.
Next up I may change the v1b cathode to be more warm bias - shouldn't affect the with feedback tone much but hoping for more hair on the without feedback.
Next up I may change the v1b cathode to be more warm bias - shouldn't affect the with feedback tone much but hoping for more hair on the without feedback.