Matchless Hotbox Layout Corrections, Need More Gain
Moderators: pompeiisneaks, Colossal
Matchless Hotbox Layout Corrections, Need More Gain
Hey there, first time posting. I recently built a Matchless Hotbox from the schematic going around the web. the schematic was OK but the layout pic was all wrong. If you built it just using the pic you wouldn't be a happy camper when you were done. Has anyone modded this for high gain. Its nice but I would just like it a lot more if it had more balls if you know what i mean. Here is the corrected pic. cheers
You do not have the required permissions to view the files attached to this post.
Last edited by dovii on Wed Aug 22, 2007 5:14 am, edited 1 time in total.
Re: Matchless Hotbox Layout Corrections, Need More Gain
Welcome D,
I contemplated combining a Hot Box and tube driven reverb into a stand along package but didn't end up building it.
Generally, gain can be increased by:
Decreasing cathode resistors
Increasing plate resistors
I start with ~5-10% changes in one place at a time and pay close attention to tube plate dissipation with each change. Also, listen closely to the tonal differences with each change to determine if you are getting closer or further away than what you desire. Changes to cathode bypass and coupling caps may help get you heading back in the direction that you desire.
Good luck,
Dave O.
I contemplated combining a Hot Box and tube driven reverb into a stand along package but didn't end up building it.
Generally, gain can be increased by:
Decreasing cathode resistors
Increasing plate resistors
I start with ~5-10% changes in one place at a time and pay close attention to tube plate dissipation with each change. Also, listen closely to the tonal differences with each change to determine if you are getting closer or further away than what you desire. Changes to cathode bypass and coupling caps may help get you heading back in the direction that you desire.
Good luck,
Dave O.
Re: Matchless Hotbox Layout Corrections, Need More Gain
Wellcome Dovii
the two 470k resistors before third stage (and the 100k after the treble pot combined with the 100k output level dimed) are 1/2 voltage divider.
so you may play with these...and get blocking distortion.
in mine i've put a 1M trim instead of the 2X470k...to be able to decrease gain.
do you send it directly in a power amp phase inverter or in an effect return? how does it sound?
i find mine a bit fuzzy sounding without singing and dynamic qualities expected (i'm more in bluesy stuff)
i've tried 100k v1 plate resistors/12ay7 in v1/12at7 in v2..but it's still not for me,so it's going to be a rocket preamp...one day
Ampgeek i'm contemplated combining a sort of TW preamp with a 6G15 reverb in a stand alone package
metalwork and cabinet done...5 months ago
the two 470k resistors before third stage (and the 100k after the treble pot combined with the 100k output level dimed) are 1/2 voltage divider.
so you may play with these...and get blocking distortion.
in mine i've put a 1M trim instead of the 2X470k...to be able to decrease gain.
do you send it directly in a power amp phase inverter or in an effect return? how does it sound?
i find mine a bit fuzzy sounding without singing and dynamic qualities expected (i'm more in bluesy stuff)
i've tried 100k v1 plate resistors/12ay7 in v1/12at7 in v2..but it's still not for me,so it's going to be a rocket preamp...one day
Ampgeek i'm contemplated combining a sort of TW preamp with a 6G15 reverb in a stand alone package
metalwork and cabinet done...5 months ago
Re: Matchless Hotbox Layout Corrections, Need More Gain
How does the math on the second one work? Got the first one. Trying to work on my tech chops vs paint-by-numbers assembler.ampgeek wrote:Welcome D,
Generally, gain can be increased by:
Decreasing cathode resistors
Increasing plate resistors
Doesn't increasing the plate resistor value increase voltage across that Rp? Wouldn't that take out more B+ and lower the plate voltage (and hence gain)?
Re: Matchless Hotbox Layout Corrections, Need More Gain
The best "math" explanation I have found (for me) is by Mr Aiken (aikenamps.com) under advanced technical articles:
"Designing Common-Cathode Triode Amplifiers"
HTH-D
"Designing Common-Cathode Triode Amplifiers"
HTH-D
Re: Matchless Hotbox Layout Corrections, Need More Gain
Hey GH,
Here is a gross oversimplification that will get your mind working in the right direction. No math needed for this one!
Think of the tube as a variable resistor passing current through the plate and, ultimately to ground. The resistance varies (e.g., is controlled by) in response to grid voltage with respect to cathode.
At static conditions (e.g., no signal at the grid), it is DC current and the voltage drop across the tube and plate resistor is fairly constant.
Now, stick an AC signal on the grid. Tube resistance (and current flow!) varies in response to the varying grid voltage differential (e.g., 440 cycles/sec for A).
Since current flow across the plate resistor is varying, the voltage drop across it "wiggles" in concert (no pun intended!) with the grid signal. Voila...DC current is "converted" into AC in response to the grid signal.
That voltage drop is proportional to the resistor value via ohms law. All else being equal, the greater the resistence value, the bigger the amplitude of the wiggle (e.g. rms voltage), the greater the gain.
Increasing the current through the tube and plate resistor does the same thing. Again, ohms law. Decreasing the cathode resistor sets the tube at a higher idle point (e.g., higher "baseline" current flow) which gives you a greater amplitude wiggle across the plate resistor for the same grid wiggle.
Hope this helps!
Dave O.
Here is a gross oversimplification that will get your mind working in the right direction. No math needed for this one!
Think of the tube as a variable resistor passing current through the plate and, ultimately to ground. The resistance varies (e.g., is controlled by) in response to grid voltage with respect to cathode.
At static conditions (e.g., no signal at the grid), it is DC current and the voltage drop across the tube and plate resistor is fairly constant.
Now, stick an AC signal on the grid. Tube resistance (and current flow!) varies in response to the varying grid voltage differential (e.g., 440 cycles/sec for A).
Since current flow across the plate resistor is varying, the voltage drop across it "wiggles" in concert (no pun intended!) with the grid signal. Voila...DC current is "converted" into AC in response to the grid signal.
That voltage drop is proportional to the resistor value via ohms law. All else being equal, the greater the resistence value, the bigger the amplitude of the wiggle (e.g. rms voltage), the greater the gain.
Increasing the current through the tube and plate resistor does the same thing. Again, ohms law. Decreasing the cathode resistor sets the tube at a higher idle point (e.g., higher "baseline" current flow) which gives you a greater amplitude wiggle across the plate resistor for the same grid wiggle.
Hope this helps!
Dave O.
Re: Matchless Hotbox Layout Corrections, Need More Gain
I've actually downloaded and printed a bunch of Aikens stuff, including that one. Have also been pounding thru NEETS. For the most part, I understand the cathode biasing schema.
What I apparently don't get is the signal path out the tube.
How is the output signal derived from the voltage across the plate resistor? I understand ohms law, etc, and that if you vary the current (AC) out the tube you then get a corresponding varying (and amplified) voltage across the plate resistor. But on one side of the Rp is the plate, with a straight path to the next stage, and the other is a B+.
I just can't seem to get out of my head that the signal should be going directly out the plate to the next stage.
What I apparently don't get is the signal path out the tube.
How is the output signal derived from the voltage across the plate resistor? I understand ohms law, etc, and that if you vary the current (AC) out the tube you then get a corresponding varying (and amplified) voltage across the plate resistor. But on one side of the Rp is the plate, with a straight path to the next stage, and the other is a B+.
I just can't seem to get out of my head that the signal should be going directly out the plate to the next stage.
Last edited by gearhead on Thu Aug 23, 2007 12:11 am, edited 2 times in total.
Re: Matchless Hotbox Layout Corrections, Need More Gain
GH,
The coupling cap in parallel with Rp and plate passess the AC (and blocks the DC!) onto the next stage.
Also note that there is an AC signal exiting the cathode when there is an AC signal on the grid. In a cathode biased arrangement, this signal is in phase with the signal on the grid while the Rp derived signal is 180 deg out of phase.
This concept is very important in understanding tube driven phase inverters (as opposed to transformer coupled PI's).
Dave O.
The coupling cap in parallel with Rp and plate passess the AC (and blocks the DC!) onto the next stage.
Also note that there is an AC signal exiting the cathode when there is an AC signal on the grid. In a cathode biased arrangement, this signal is in phase with the signal on the grid while the Rp derived signal is 180 deg out of phase.
This concept is very important in understanding tube driven phase inverters (as opposed to transformer coupled PI's).
Dave O.
Re: Matchless Hotbox Layout Corrections, Need More Gain
Grack, still don't get it. I can see from formula
Voltage Gain = Rp*mu/(Rp+ra).
Mathwise, yepp, increase Rp and you increase voltage gain. I just can't visualize why the output is a parallel of Rp and ra.
WRT to the phase, understand the cathode AC being in-phase with the grid; as grid signal goes positive, it decreases the negativeness of the bias (less negative, more positive LOL). This decreases the suppressive effects of the grid with the plate able to suck more electrons from the cathode.
WRT the Rp-based signal out of phase, got me there. The only way I can see the Rp-based output being out of phase is that as the current increases, voltage across Rp increases, which would lower the plate voltage? That then assumes the predominant signal is out the plate. A close example of this is in NEETS, module 6, page 1-25
Thanks for your help!
Voltage Gain = Rp*mu/(Rp+ra).
Mathwise, yepp, increase Rp and you increase voltage gain. I just can't visualize why the output is a parallel of Rp and ra.
WRT to the phase, understand the cathode AC being in-phase with the grid; as grid signal goes positive, it decreases the negativeness of the bias (less negative, more positive LOL). This decreases the suppressive effects of the grid with the plate able to suck more electrons from the cathode.
WRT the Rp-based signal out of phase, got me there. The only way I can see the Rp-based output being out of phase is that as the current increases, voltage across Rp increases, which would lower the plate voltage? That then assumes the predominant signal is out the plate. A close example of this is in NEETS, module 6, page 1-25
Thanks for your help!
Re: Matchless Hotbox Layout Corrections, Need More Gain
> How is the output signal derived from the voltage across the plate resistor?
> on one side of the Rp is the plate, with a straight path to the next stage, and the other is a B+.
100K resistor. 1mA standing current. 100V across the resistor.
Apply a signal. Wobble the current +/-0.1mA, or 1.1mA - 0.9mA.
Resistor voltage wobbles 110V - 90V.
> the other is a B+.
And where is that B+ going? Nowhere! It is stuck at +242V (or whatever).
So if one end of the resistor is nailed to 242V, and the resistor drop wobbles 110V - 90V, then the other end of the resistor must wobble 132V - 152V.
We have a 20V wobble added to a 142V standing voltage.
At turn-on, the coupling cap to the next stage charges-up to the difference between that 142V and the voltage at the next stage (usually zero in tubework). After that, the coupling cap pretty-much holds its voltage. So at the next-grid end of the cap we find zero steady voltage and our 20V audio wobble.
> out of phase, got me there.
> current increases, voltage across Rp increases, which would lower the plate voltage
Right.
And when looking at Mu.... Mu is always a negative number (inverting), but we neglect to note that in the datasheet or run it through formulas because we already know that part of the final answer ("it inverts").
As for the general question: Why does it invert? It just does. In fact all single-device amplifiers with both current gain and voltage gain, using vacuum tubes or bipolar transistors or FETs, DO invert voltage.
This reflects the fact that at the individual charge-particle level, all these devices work the same. The external field imposed by the vacuum or crystal affects whether the sum of all the individual charge-particle actions gives infinite voltage gain, infinite current gain, or both or neither. (Gain is never infinite. But there's a difference between the Mu=8 of an old triode and Mu=500+ of a pentode or BJT. The current gain of a vacuum pentode at DC is over a million, and still over 1,000 at the top of the audio band, the current gain of a BJT or positive-grid vacuum tube is 10-300. When gain is around 10, it matters. When gain is many hundred, it usually does not limit useful design.)
Note that current gain in these stages is non-inverting. It is hard to show on small vacuum tubes (and FETS) because we stay away from grid current. It is quite obvious in BJT amps: more base current is more collector current.
And for further amusement: common-plate (cathode follower) and common-grid stages do not invert voltage. They also have unity gain of either current or voltage.
My wild-guess is that any 3-terminal device which did not invert voltage yet had both voltage gain and current gain would instantly force itself full-on or full-off, and never be good for a linear amplifier. But I can't find an example.
You actually do not need to grok this stuff to pencil-sketch the HotBox. Find the Resistance-Coupled Amplifier tables. Changing Rp from 100K to 220K, or Rk from 1.5K to 2.2K, is piddly little 10%-20% changes of gain. With cathode cap and not-low loading, 12AX7 will give gain of 40-55. You ballpark by penciling "50?" on each stage. Without cathode cap, gain is around half, pencil "25?" by each stage.
This assumes that the cathode cap is "large" at all frequencies of interest. To be very sure, take Rk and divide by 2 (for all 12AX7 work, "500 ohms" is extra-safe), pick a frequency far below your lowest tone (20Hz for hi-fi, 80Hz for guitar), and find what cap has that impedance. 500 ohms at 20Hz is 20uFd, which is indeed a popular cathode cap for full-bass small-bottle work.
Here we find cathode caps 20 to 200 times smaller than we'd pick for extra-full 20Hz response. These stages have gain of 50 at high frequency and 25 at low frequency. The "split" between "high" and "low" happens at 200Hz or 2KHz, depending on the cap (and a bit on tube bias). The amp is un-flat, and these cap values were surely picked by ear for a particular sound. I don't see much point in computing the precise frequency response. We just know that the trend is 25 per stage in bass, 50 per stage in treble, this boost is common in guitar, and the inflection point can be swiftly changed and tried "by ear" with a cup full of caps and a hot iron.
So using numbers like 25 or 50 per stage, the 2:1 loss in the 470K resistors that klingo notes, and assuming 20:1 or 30:1 loss in the tone-stack.... you can show that when both pots are full-up, and driving an INST input, this plan is NOT low-gain. So why is dovii asking for more?
> on one side of the Rp is the plate, with a straight path to the next stage, and the other is a B+.
100K resistor. 1mA standing current. 100V across the resistor.
Apply a signal. Wobble the current +/-0.1mA, or 1.1mA - 0.9mA.
Resistor voltage wobbles 110V - 90V.
> the other is a B+.
And where is that B+ going? Nowhere! It is stuck at +242V (or whatever).
So if one end of the resistor is nailed to 242V, and the resistor drop wobbles 110V - 90V, then the other end of the resistor must wobble 132V - 152V.
We have a 20V wobble added to a 142V standing voltage.
At turn-on, the coupling cap to the next stage charges-up to the difference between that 142V and the voltage at the next stage (usually zero in tubework). After that, the coupling cap pretty-much holds its voltage. So at the next-grid end of the cap we find zero steady voltage and our 20V audio wobble.
> out of phase, got me there.
> current increases, voltage across Rp increases, which would lower the plate voltage
Right.
And when looking at Mu.... Mu is always a negative number (inverting), but we neglect to note that in the datasheet or run it through formulas because we already know that part of the final answer ("it inverts").
As for the general question: Why does it invert? It just does. In fact all single-device amplifiers with both current gain and voltage gain, using vacuum tubes or bipolar transistors or FETs, DO invert voltage.
This reflects the fact that at the individual charge-particle level, all these devices work the same. The external field imposed by the vacuum or crystal affects whether the sum of all the individual charge-particle actions gives infinite voltage gain, infinite current gain, or both or neither. (Gain is never infinite. But there's a difference between the Mu=8 of an old triode and Mu=500+ of a pentode or BJT. The current gain of a vacuum pentode at DC is over a million, and still over 1,000 at the top of the audio band, the current gain of a BJT or positive-grid vacuum tube is 10-300. When gain is around 10, it matters. When gain is many hundred, it usually does not limit useful design.)
Note that current gain in these stages is non-inverting. It is hard to show on small vacuum tubes (and FETS) because we stay away from grid current. It is quite obvious in BJT amps: more base current is more collector current.
And for further amusement: common-plate (cathode follower) and common-grid stages do not invert voltage. They also have unity gain of either current or voltage.
My wild-guess is that any 3-terminal device which did not invert voltage yet had both voltage gain and current gain would instantly force itself full-on or full-off, and never be good for a linear amplifier. But I can't find an example.
You actually do not need to grok this stuff to pencil-sketch the HotBox. Find the Resistance-Coupled Amplifier tables. Changing Rp from 100K to 220K, or Rk from 1.5K to 2.2K, is piddly little 10%-20% changes of gain. With cathode cap and not-low loading, 12AX7 will give gain of 40-55. You ballpark by penciling "50?" on each stage. Without cathode cap, gain is around half, pencil "25?" by each stage.
This assumes that the cathode cap is "large" at all frequencies of interest. To be very sure, take Rk and divide by 2 (for all 12AX7 work, "500 ohms" is extra-safe), pick a frequency far below your lowest tone (20Hz for hi-fi, 80Hz for guitar), and find what cap has that impedance. 500 ohms at 20Hz is 20uFd, which is indeed a popular cathode cap for full-bass small-bottle work.
Here we find cathode caps 20 to 200 times smaller than we'd pick for extra-full 20Hz response. These stages have gain of 50 at high frequency and 25 at low frequency. The "split" between "high" and "low" happens at 200Hz or 2KHz, depending on the cap (and a bit on tube bias). The amp is un-flat, and these cap values were surely picked by ear for a particular sound. I don't see much point in computing the precise frequency response. We just know that the trend is 25 per stage in bass, 50 per stage in treble, this boost is common in guitar, and the inflection point can be swiftly changed and tried "by ear" with a cup full of caps and a hot iron.
So using numbers like 25 or 50 per stage, the 2:1 loss in the 470K resistors that klingo notes, and assuming 20:1 or 30:1 loss in the tone-stack.... you can show that when both pots are full-up, and driving an INST input, this plan is NOT low-gain. So why is dovii asking for more?
I'm such a butt head, found lots of gain!!
Hey guys thanks for your replies. You know I have always found that if you go back enough times (6 or 7) you find out how you screwed it up. My problem was at the very first stage cathode resistor, should be 2.2k and I had 220k in its place. D'oh Must have had too many happy pops that night. I have built lots of amps (300b SE, 18watt trem, Dynacos etc..) but this project has been a little tyrant.
Re: Matchless Hotbox Layout Corrections, Need More Gain
Whoo hooo! Thanks, that makes sense.
Even better for this simplistic example,
If you INCREASE the plate load resistor to, say, 120K, then the plate is at 122V (242V-120k*1ma) and:
Apply a signal. Wobble the current +/-0.1mA, or 1.1mA - 0.9mA.
Resistor voltage wobbles 132V - 108V.
> the other is a B+.
And where is that B+ going? Nowhere! It is stuck at +242V (or whatever).
So if one end of the resistor is nailed to 242V, and the resistor drop wobbles 132V - 108V, then the other end of the resistor must wobble 110V - 134V
We have a 24V wobble added to a 122V standing voltage.
100k increase to 120k results in increase output from 20V to 24V.
Now, I assume this is a simplistic example, and doesn't take into account the effects of the tube (resistance, capactiance, current limiting, etc), so the reuslt isn't as linear as this. But it does explain, from a proportional standpoint, why an INCREASE in plate load resistor increases voltage gain.
Also btw, this example (to me) explains the inversion. As the (AC) wiggling current increases out of the plate, the voltage across the plate load resistor increases. This in turn drives -down- the voltage at the bottom of the plate load resistor, where your output is taken. Current Up>>Voltage Down; direct out-of-phase relationship.
By the way, where are the tables you mention in many of your posts? Have been furiously downloading texts available online, and haven't come across most of the ones you refererence.
Even better for this simplistic example,
If you INCREASE the plate load resistor to, say, 120K, then the plate is at 122V (242V-120k*1ma) and:
Apply a signal. Wobble the current +/-0.1mA, or 1.1mA - 0.9mA.
Resistor voltage wobbles 132V - 108V.
> the other is a B+.
And where is that B+ going? Nowhere! It is stuck at +242V (or whatever).
So if one end of the resistor is nailed to 242V, and the resistor drop wobbles 132V - 108V, then the other end of the resistor must wobble 110V - 134V
We have a 24V wobble added to a 122V standing voltage.
100k increase to 120k results in increase output from 20V to 24V.
Now, I assume this is a simplistic example, and doesn't take into account the effects of the tube (resistance, capactiance, current limiting, etc), so the reuslt isn't as linear as this. But it does explain, from a proportional standpoint, why an INCREASE in plate load resistor increases voltage gain.
Also btw, this example (to me) explains the inversion. As the (AC) wiggling current increases out of the plate, the voltage across the plate load resistor increases. This in turn drives -down- the voltage at the bottom of the plate load resistor, where your output is taken. Current Up>>Voltage Down; direct out-of-phase relationship.
By the way, where are the tables you mention in many of your posts? Have been furiously downloading texts available online, and haven't come across most of the ones you refererence.
-
skyboltone
- Posts: 2287
- Joined: Wed May 10, 2006 7:02 pm
- Location: Sparks, NV, where nowhere looks like home.
Re: Matchless Hotbox Layout Corrections, Need More Gain
Get an RCA tube manual. Try Amazon's used books. The tables are in the back of the book for resistance coupled amplifiers. They give common values for components used around about a dozen tubes and their variants.gearhead wrote:By the way, where are the tables you mention in many of your posts? Have been furiously downloading texts available online, and haven't come across most of the ones you reference.
Dan H
The Last of the World's Great Human Beings
Seek immediate medical attention if you suddenly go either deaf or blind.
If you put the Federal Government in charge of the Sahara Desert, in five years time there would be a shortage of sand.
Seek immediate medical attention if you suddenly go either deaf or blind.
If you put the Federal Government in charge of the Sahara Desert, in five years time there would be a shortage of sand.
Re: Matchless Hotbox Layout Corrections, Need More Gain
> where are the tables you mention in many of your posts?
Many collectors don't realize that the R-C Amplifier Tables are THE designer's crutch.
Pete Millett has the Sylvania tube manuals posted, but has the R-C lumped in with other end-matter.
Here is Sylvania 1959 Resistance-Coupled Amplifier Tables. 1.6MB PDF.
Don't skip the first 3 pages!
Many collectors don't realize that the R-C Amplifier Tables are THE designer's crutch.
Pete Millett has the Sylvania tube manuals posted, but has the R-C lumped in with other end-matter.
Here is Sylvania 1959 Resistance-Coupled Amplifier Tables. 1.6MB PDF.
Don't skip the first 3 pages!
You do not have the required permissions to view the files attached to this post.
Re: Matchless Hotbox Layout Corrections, Need More Gain
PRR - Thanks once again! After you mentioned where it was, I also found out where Millett buried those ever-so-useful pages.
SB - Do you know if the commonly reprinted RCA book (RC-30, 1973) has R-C tables in the back for the EL-34/6CA7 and 6L6? Earlier RCA versions available on-line have data sheets for those, but not R-C tables (do for the 12AX7).
SB - Do you know if the commonly reprinted RCA book (RC-30, 1973) has R-C tables in the back for the EL-34/6CA7 and 6L6? Earlier RCA versions available on-line have data sheets for those, but not R-C tables (do for the 12AX7).