One more question (for now) 
I read a ton and keep piles of notes and thoughts I've compiled from the forums before I decide that I'm actually ready for a build. I found a note to consider powering the BA3b front-end boards with a different PSU - possibly incorporating the super-regulator - than the output stage. I scanned back through the thread, and I can't find where I got this idea. I know I didn't just pull it from thin air, but I can never be sure.
Is that absurd? I'd be curious to hear some thoughts.
I read a ton and keep piles of notes and thoughts I've compiled from the forums before I decide that I'm actually ready for a build. I found a note to consider powering the BA3b front-end boards with a different PSU - possibly incorporating the super-regulator - than the output stage. I scanned back through the thread, and I can't find where I got this idea. I know I didn't just pull it from thin air, but I can never be sure.
Is that absurd? I'd be curious to hear some thoughts.
To be clear, are you planning on a power amp or a pre-amp?
If the latter, then it's entirely sensible to me that, because a stand-alone pre- is a separate bit of kit, it would be developed with its own power supply by default. If it's part of a power amp, then I'd think it's a question of personal preference and available real estate inside the chassis you intend to use.
From http://www.firstwatt.com/pdf/art_mos_test.pdf
Why do you want to use a lower resistance in the first place? That's the bit that you haven't explained (I don't think) to those lacking knowledge like me. I dare say that there'll be more than one fearless amp building guru who'll jump in to clarify that point for me because it's patently obvious to those with the insight. I value the education, so don't mind asking daft questions.
Interested to read the follow up..
If the latter, then it's entirely sensible to me that, because a stand-alone pre- is a separate bit of kit, it would be developed with its own power supply by default. If it's part of a power amp, then I'd think it's a question of personal preference and available real estate inside the chassis you intend to use.
From http://www.firstwatt.com/pdf/art_mos_test.pdf
My read of the above, and it's probably overly simplistic, is that selecting a resistance of 1 Ω makes the maths simple when applying Ohms Law of V = I*R.
Why do you want to use a lower resistance in the first place? That's the bit that you haven't explained (I don't think) to those lacking knowledge like me. I dare say that there'll be more than one fearless amp building guru who'll jump in to clarify that point for me because it's patently obvious to those with the insight. I value the education, so don't mind asking daft questions.
Interested to read the follow up..
Last edited:
Hi Astromo -
Sorry about the nomenclature. I'm using the "modules" for a few power amp iterations. In the articles, I noticed that I can mix and match so to speak.
First - I'm planning on using the BA3 front end with the BA-2 complimentary output stages for stereo amp.
Second - I'll convert the stereo amp to a monoblock and fire up another monoblock. Along with this, I'll hope to have a plan for making them switchable between stereo and balanced mono. I've seen the excellent PCB for the switching on the BA3b boards and a few solutions.
Third - I'll be swapping out the output stages to the BA-1 SE versions in one chassis to compare SE to Complimentary using the same BA3 front end. When it's all said and done, there will also be a few experiments with bias and number of output devices tossed in for good measure. I finally have an oscilloscope; a way to measure distortion along with H2 etc. will be on the way, and I'm curious.
------------
Onto the why for the source resistors - I had planned on using the 1R0 in the schematic. It was suggested to me to lower the value to 0R22. The suggestion included an explanation that flew a bit over my head, but in essence I think it was less voltage loss at the output with no penalty (except matching) which means more potential power. I had also read prior that the best source resistor is none at all, but stability can be an issue. That makes sense.
So, if I can swap out the resistors for common ones I already have in a bag (I have 100s of 0R22 and 0R47, but no 1R0) with no penalty and a potential benefit - then why not.
I trust the source of the suggestion implicitly, but I wanted to see if I could determine why it would work and how closely I'd need to match the devices to be sure I'd be okay and learn a bit along the way. This will definitely come in handy for future builds.
Also, I'm planning on matching at roughly 0A5 and 24V for the first time using a slightly different process than for the First Watt amps I built. These devices may also see a bit higher bias than some of my First Watt builds. So, I also wanted to make sure I had some padding in the numbers. If the math says Vgs should be within 0V005 for the safe use of 0R22 and 0V05 for 0R47 as an example, I'd definitely go with the 0R47.
I hope that makes sense.
Sorry about the nomenclature. I'm using the "modules" for a few power amp iterations. In the articles, I noticed that I can mix and match so to speak.
First - I'm planning on using the BA3 front end with the BA-2 complimentary output stages for stereo amp.
Second - I'll convert the stereo amp to a monoblock and fire up another monoblock. Along with this, I'll hope to have a plan for making them switchable between stereo and balanced mono. I've seen the excellent PCB for the switching on the BA3b boards and a few solutions.
Third - I'll be swapping out the output stages to the BA-1 SE versions in one chassis to compare SE to Complimentary using the same BA3 front end. When it's all said and done, there will also be a few experiments with bias and number of output devices tossed in for good measure. I finally have an oscilloscope; a way to measure distortion along with H2 etc. will be on the way, and I'm curious.
------------
Onto the why for the source resistors - I had planned on using the 1R0 in the schematic. It was suggested to me to lower the value to 0R22. The suggestion included an explanation that flew a bit over my head, but in essence I think it was less voltage loss at the output with no penalty (except matching) which means more potential power. I had also read prior that the best source resistor is none at all, but stability can be an issue. That makes sense.
So, if I can swap out the resistors for common ones I already have in a bag (I have 100s of 0R22 and 0R47, but no 1R0) with no penalty and a potential benefit - then why not.
I trust the source of the suggestion implicitly, but I wanted to see if I could determine why it would work and how closely I'd need to match the devices to be sure I'd be okay and learn a bit along the way. This will definitely come in handy for future builds.
Also, I'm planning on matching at roughly 0A5 and 24V for the first time using a slightly different process than for the First Watt amps I built. These devices may also see a bit higher bias than some of my First Watt builds. So, I also wanted to make sure I had some padding in the numbers. If the math says Vgs should be within 0V005 for the safe use of 0R22 and 0V05 for 0R47 as an example, I'd definitely go with the 0R47.
I hope that makes sense.
Thank you, ZM!!!For everyone's benefit - because I know I can't be the ONLY person that wonders about these things - here is a summary of the Mighty ZM's notes and my personal decision.
1) The choice in the value of the source resistor is not only dependent on matching / sorting the Vgs of the output devices, but on the type of bias circuit used. This makes perfect sense, but I never considered it.
2) For the bias circuit in the BA-2 - there is no compensation mechanism for the tempco.
3) We did not go through the math. Sticking strictly to shared practical experience. I can learn that later.
My own decision is to go ahead and get the matching process completed. Then, I will choose between 0R47 or 1R0. In matching two tubes of the 240s and 9240s in the past, I got very tight values. However, I'm using a different process this time at higher current, so we'll see how it turns out.
When you say "BA3 front end mixed and matched to BA-2 Complementary output," you mean you are building a BA3 amp, right? Same boards for complementary BA2 and BA3.
I only ask because I thought maybe I missed something, which is way possible!
Russellc
Hi Russellc -
I doubt you missed anything. I probably did. Short answer is: yes. I am starting with the "BA-3". I want to learn to tweak P3 and then compare it to another amp or two.
If I like it, then that specific build will be instantly converted (either flip a few switches or swap some wires) to a BA3b monoblock, and I'll add the other monoblock. I have all the boards etc. I'm just moving through a slow progression with some fun experiments along the way.
FWIW - I am just not sure what to call each thing, so I'm describing the modules. Here's the way I understand it per the .pdf and the articles.
BA1 amplifier => BA1 Front End + BA1 Single Ended Output Stage
BA2 amplifier => BA2 Front End + BA2 Complimentary Output Stage
BA3 amplifier => BA3 Front End + BA2 Complimentary Output Stage
BA? amplifier => BA3 Front End + BA1 Single Ended Output Stage
BA3b amplifier => Balanced BA3
BA?b amplifier => Balanced BA?
I plan on trying all but the BA-1 and BA-2, but that could change.
I doubt you missed anything. I probably did.
Short answer is: yes. I am starting with the "BA-3". I want to learn to tweak P3 and then compare it to another amp or two. If I like it, then that specific build will be instantly converted (either flip a few switches or swap some wires) to a BA3b monoblock, and I'll add the other monoblock. I have all the boards etc. I'm just moving through a slow progression with some fun experiments along the way.
FWIW - I am just not sure what to call each thing, so I'm describing the modules. Here's the way I understand it per the .pdf and the articles.
BA1 amplifier => BA1 Front End + BA1 Single Ended Output Stage
BA2 amplifier => BA2 Front End + BA2 Complimentary Output Stage
BA3 amplifier => BA3 Front End + BA2 Complimentary Output Stage
BA? amplifier => BA3 Front End + BA1 Single Ended Output Stage
BA3b amplifier => Balanced BA3
BA?b amplifier => Balanced BA?
I plan on trying all but the BA-1 and BA-2, but that could change.
Last edited:
One more question (for now)
I read a ton and keep piles of notes and thoughts I've compiled from the forums before I decide that I'm actually ready for a build. I found a note to consider powering the BA3b front-end boards with a different PSU - possibly incorporating the super-regulator - than the output stage. I scanned back through the thread, and I can't find where I got this idea. I know I didn't just pull it from thin air, but I can never be sure.
Is that absurd? I'd be curious to hear some thoughts.
I remember someone who did this as well, regulated the BA3 FE separately in a BA3 Power Amplifier.
Russellc
Hi Russellc -
I doubt you missed anything. I probably did.
If I like it, then that specific build will be instantly converted (either flip a few switches or swap some wires) to a BA3b monoblock, and I'll add the other monoblock. I have all the boards etc. I'm just moving through a slow progression with some fun experiments along the way.
FWIW - I am just not sure what to call each thing, so I'm describing the modules. Here's the way I understand it per the .pdf and the articles.
BA1 amplifier => BA1 Front End + BA1 Single Ended Output Stage
BA2 amplifier => BA2 Front End + BA2 Complimentary Output Stage
BA3 amplifier => BA3 Front End + BA2 Complimentary Output Stage
BA? amplifier => BA3 Front End + BA1 Single Ended Output Stage
BA3b amplifier => Balanced BA3
BA?b amplifier => Balanced BA?
I plan on trying all but the BA-1 and BA-2, but that could change.
I'm in the same boat at a similar place with P3. I started a thread some time back and STILL haven't set up focusrite yet.
Similarly as with BA3 complementary vs BA2 OS, Single ended is same way except it will also work, of course with BA1. Somewhere in here someone had to point out to me BA1 and BA2 FE board is same board. So many Pass Clone amps, so little time!
I at least at this time, do not have an interest in BA1,BA2 or BA3 single ended. I started some long while ago getting interested in complementary BA2, (have partially stuffed board around here somewhere) and then BA3 came on the scene and that did it! That has lead me to BBA3 mono blocks, which I have collected all parts needed except two 5U deluxe chassis and transformers.
Once Iron pre, DCG3 and Sissysit amp done, focus will begin in earnest. In mean time I need to get Focusrite and P3 adjustment down.
Russellc
We must be brothers from different mothers
I just got the JFETs for the Balanced Iron Pre, BA, and other future builds along with some extra Keratherm in the mail today. That leaves only matching the output devices and getting some re-jiggering of the chassis done. The steel cans from Antek were a bit bigger than I thought. I hope to have everything by next weekend.
My Focusrite should be here soon. I read the REW thread a number of times, but until I get my hands on it and play - I won't know how successful I'll be.
I just scored a second-hand oscilloscope to try and learn a few other basic measurements. The probes should be here in the next day or two. That's a manual I don't look forward to reading. I read the "quick guide", and that was enough to cross my eyes.
Love to compare notes along the way - I learn a ton from everyone in the forums!
I just got the JFETs for the Balanced Iron Pre, BA, and other future builds along with some extra Keratherm in the mail today. That leaves only matching the output devices and getting some re-jiggering of the chassis done. The steel cans from Antek were a bit bigger than I thought. I hope to have everything by next weekend.
My Focusrite should be here soon. I read the REW thread a number of times, but until I get my hands on it and play - I won't know how successful I'll be.
I just scored a second-hand oscilloscope to try and learn a few other basic measurements. The probes should be here in the next day or two. That's a manual I don't look forward to reading. I read the "quick guide", and that was enough to cross my eyes.
Love to compare notes along the way - I learn a ton from everyone in the forums!
Just happens to be my favorite color. Is it wrong to pick a cap based on color?
Reminds me of days past 426 Hemi Challenger/Charger purple!
Russellc
View attachment 838207
vs.
View attachment 838208
AND
View attachment 838210
vs.
View attachment 838209
Correlation may not be causation, but coincidence? I think not
The Dart is the color purple I referred to. You could get 340 six pack as I recall. The 426 Hemi was a beast. Advertised at 425 H.P. and about the same torque. They were way brutal. I knew the manufacturer rated them conservatively, but didnt know how much. They were in huge cars that were brutally fast at drag strip.
In recent years, I saw an automotive program where they reproduced a 426 Hemi from stock parts, (yes it would be very, very expensive to do so) and dyno tested it. Over 800HP. Now I know why those cars were so brutal.
Russellc
Before I start matching my devices tomorrow for my BA builds, I’d love to get a better understanding around matching. I put this here for visibility and b/c it does directly relate to building the BA-1 and BA-2 output stages for use in conjunction with the BA-3 front end. Making a BA-3b (to start). If it needs to be moved... please feel free.
This topic has probably been beaten to death over the years, but I can’t find one common thread or even a few sources I can digest for matching as it relates to Vgs. Those that know me, know I read a ton and make every attempt to digest things before asking.
I’ve successfully executed the matching process a number of times for First Watt clones and those of a few friends, basically playing the roll of bench tech. However, I’m still not completely sorted on the fundamentals of “why” except to repeat by rote two phrases I’ve commonly seen / heard:
Assume that I am referencing devices in parallel for output stages like the BA-1 and BA-2.
I’ve read the two Pass articles a number of times, which are phenomenal, but I’m still not “clicking”.
For Number 1: In a design where we have a “high” Class A bias current, I assumed quite possibly incorrectly that all devices were always on. Thus the lower efficiency compared to other types. So, how does #1 apply? If we go to the water analogy; I thought that the spigot (or all smaller spigots funneling into a larger pipe) were always “open”. However, I could imagine in my mind if 6 devices in parallel had one device turn on “early” and all the current required were trying to go through that single device even instantaneously – that could be a failure event. If that one device wasn’t rated for the total current, it might go kaput or be outside of it's happy place. But if it were a big ol’ honking puck that could handle all the current … then what’s the point of paralleling the devices? Is it to find that “sweet spot” and keep all the devices in their happy place rather than just cranking the snot out of one or two? Is the “always on” analogy only relevant to Single Ended designs? I thought for push-pull / complimentary that one half of the signal was handled by the P’s and the other the N’s, but that all devices were still on / ready and waiting for the “handoff” so to speak.
Number 2: This concept seems to click a bit more when I read about transconductance and its potential variation. Example, the devices “match” for Vgs at a current of 0A25, but not at 0A50. I am not fully in tune with transconductance, but it seems to be a delta/ratio/metric for describing the variable non-linearity of current flow vs. voltage. If one device of the 6 had a transconductance of 1.5 and all others were at 1.0 – unless compensated with a source resistor (or source resistors) – then that device would be “working a lot harder”, flowing a lot more current, get hotter etc. than the 5 others. Not ideal. I’m still not certain how to calculate a ballpark source resistance common across all parallel devices, but that can come later. I trust the schematics and the rules. I’m trying to grasp the basics. I know “how” to match per following the recipe, but I’d love to learn more.
tl;dr – are we really trying to match for transconductance by measuring Vgs even at only one current and generally assuming that the source resistors will take up the slack for some of the variability? Would it be better to match for transconductance? i.e. measure at say two or more currents? Is that the concept of curve tracing?
Be gentle… old dog… new tricks and all that
This topic has probably been beaten to death over the years, but I can’t find one common thread or even a few sources I can digest for matching as it relates to Vgs. Those that know me, know I read a ton and make every attempt to digest things before asking.
I’ve successfully executed the matching process a number of times for First Watt clones and those of a few friends, basically playing the roll of bench tech. However, I’m still not completely sorted on the fundamentals of “why” except to repeat by rote two phrases I’ve commonly seen / heard:
1) To ensure the MOSFETs turn on at as close to the same “time” as possible.
2) To ensure the MOSFETs share the current as equally as possible.
Those two seem very related, but also not precisely the same.
2) To ensure the MOSFETs share the current as equally as possible.
Those two seem very related, but also not precisely the same.
Assume that I am referencing devices in parallel for output stages like the BA-1 and BA-2.
I’ve read the two Pass articles a number of times, which are phenomenal, but I’m still not “clicking”.
For Number 1: In a design where we have a “high” Class A bias current, I assumed quite possibly incorrectly that all devices were always on. Thus the lower efficiency compared to other types. So, how does #1 apply? If we go to the water analogy; I thought that the spigot (or all smaller spigots funneling into a larger pipe) were always “open”. However, I could imagine in my mind if 6 devices in parallel had one device turn on “early” and all the current required were trying to go through that single device even instantaneously – that could be a failure event. If that one device wasn’t rated for the total current, it might go kaput or be outside of it's happy place. But if it were a big ol’ honking puck that could handle all the current … then what’s the point of paralleling the devices? Is it to find that “sweet spot” and keep all the devices in their happy place rather than just cranking the snot out of one or two? Is the “always on” analogy only relevant to Single Ended designs? I thought for push-pull / complimentary that one half of the signal was handled by the P’s and the other the N’s, but that all devices were still on / ready and waiting for the “handoff” so to speak.
Number 2: This concept seems to click a bit more when I read about transconductance and its potential variation. Example, the devices “match” for Vgs at a current of 0A25, but not at 0A50. I am not fully in tune with transconductance, but it seems to be a delta/ratio/metric for describing the variable non-linearity of current flow vs. voltage. If one device of the 6 had a transconductance of 1.5 and all others were at 1.0 – unless compensated with a source resistor (or source resistors) – then that device would be “working a lot harder”, flowing a lot more current, get hotter etc. than the 5 others. Not ideal. I’m still not certain how to calculate a ballpark source resistance common across all parallel devices, but that can come later. I trust the schematics and the rules. I’m trying to grasp the basics. I know “how” to match per following the recipe, but I’d love to learn more.
tl;dr – are we really trying to match for transconductance by measuring Vgs even at only one current and generally assuming that the source resistors will take up the slack for some of the variability? Would it be better to match for transconductance? i.e. measure at say two or more currents? Is that the concept of curve tracing?
Be gentle… old dog… new tricks and all that
Last edited: