Need PA amp conversion help

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

PRR

Member
Joined 2003
Paid Member
...The omission of any y-axis scale in the online calculator seems inexplicable, though.

It is. I may drop Richard a note.

It seems that if the deviation is "significant", you get scale. However the values we are trying here give only 23.33-23.90 dB spread, less than a dB. The scale seems to be 2dB increments. 0.54dB blows-up so big that the +/-2dB marks are off-graph. Once you realize this unexpected action, you know that "no marks" means "no effect" within the accuracy of guitar amp work which is full of 6dB and 10dB bumps.
 

Attachments

  • AmpBooks-CCalc-12AX7.gif
    AmpBooks-CCalc-12AX7.gif
    28.1 KB · Views: 57
The scale seems to be 2dB increments.
If frozen at 2 dB, that is a bad idea. It should re-scale the vertical axis based on maximum and minimum values of the plot.

Because of the shape of the shelving-filter response, this is child's play in this case, because all you have to do is find the voltage gain at the lowest frequency (10 Hz), which is automatically the lowest point on the plot, and the gain at the highest frequency (10 kHz), which is automatically the highest point on the plot. No need to hunt through every data point looking for maxima and minima, as you'd normally have to do.

If the intention is to alert the user that there isn't enough gain change to matter, that should be a separate message, preferably overlaid on the graph.

If you do contact Mr. Kuehnel , perhaps you could pass on those suggestions?

About thirty years ago, I got bored one summer, and spent a couple of weeks writing my own graph-plotting program using my housemate's 386 PC and either Quick C or Turbo C, I forget which. Figuring out how to auto-scale the y-axis, and only generate sane (human-friendly) values for the y-axis "ticks" was a challenge, but I cracked it.

As an aside, LTSpice is ridiculously awful at this. It auto-scales, but comes up with things like 0.783 dB per tick on the vertical axis. Utterly ridiculous, and IMO the single worst feature of the software.


-Gnobuddy
 
I will probably have to remedy this situation and buy some scoping equipment sometime...
In the US, you can sometimes get lucky and find an old analogue scope in good working condition for less than $100.

When I moved to BC from Los Angeles, I was shocked to find that the only used 'scopes I could find around here cost $600 CAD. Perhaps I was looking in the wrong places.

At any rate, to me, that was silly money for equipment that cost $700 new in 1990, and had since been through 25 years of unknown service and calibration history.

I ended up with a brand new 4-channel Rigol for around $450 USD. It can do many useful digital tricks, but the crappy 7-bit vertical resolution of all the affordable digital scopes is frustratingly bad. Every trace is jaggy, literally even worse than the stair-step graphic edges we put up with in the 1980s with our shiny new 320x240 computer monitors.
I think the amp is close to being usable
I hope you get it there soon.

And when you do, you may find yourself standing at the top of a very long, very slippery, very steep slope - the one that leads to endless amp-building and tinkering, while the hours you used to dedicate to guitar practice are diverted to cutting wires and soldering joints. :no:

Then one day you look around and see seven half-finished guitar amps, but the calluses on your fretting hand have disappeared, and you've forgotten how to play a Bbmaj7. :boggled:

And you tell yourself that just as soon as you finally build the perfect amp, you'll get right back to the guitar, and sound better than ever. So you head back to the soldering iron. :hypno1:

I think there's a country song waiting to be found somewhere in that narrative...


-Gnobuddy
 
I
I ended up with a brand new 4-channel Rigol for around $450 USD. It can do many useful digital tricks, but the crappy 7-bit vertical resolution of all the affordable digital scopes is frustratingly bad. Every trace is jaggy, literally even worse than the stair-step graphic edges we put up with in the 1980s with our shiny new 320x240 computer monitors.
I've been pretty darn impressed with the Keysight (nee Aglient nee HP) 1000 X-series CROs, which start from about US$500 and are built on custom ASICs. The 1000 series are WVGA (even the up to $20K 4000 series are too - but they've some graphics smarts to help too )

Haven't bought one yet as I don't have a "burning platform" at the office and I've already got two scopes at home :D
 
Last edited:
In the US, you can sometimes get lucky and find an old analogue scope in good working condition for less than $100.
Something like this could be ideal if I can find it.:up:

I ended up with a brand new 4-channel Rigol for around $450 USD. It can do many useful digital tricks, but the crappy 7-bit vertical resolution of all the affordable digital scopes is frustratingly bad. Every trace is jaggy, literally even worse than the stair-step graphic edges we put up with in the 1980s with our shiny new 320x240 computer monitors.
Interesting trade off lol. I don't think I'll need something capable of "digital tricks" for my new hobby just yet, but on that note...

And when you do, you may find yourself standing at the top of a very long, very slippery, very steep slope - the one that leads to endless amp-building and tinkering, while the hours you used to dedicate to guitar practice are diverted to cutting wires and soldering joints.

Then one day you look around and see seven half-finished guitar amps, but the calluses on your fretting hand have disappeared, and you've forgotten how to play a Bbmaj7.

And you tell yourself that just as soon as you finally build the perfect amp, you'll get right back to the guitar, and sound better than ever. So you head back to the soldering iron.
BAHAHAHA :rofl: you may have missed your calling in writing good drama! Fortunately I don't think my wife and kids will allow me enough free time to get in that deep...
However, I did find myself buying a used pedal yesterday for no other reason than to have something very small to tinker with at home. Modding and swapping out capacitors and resistors in the circuit to change and perfect the sound. Soldering my evening away at the dinner table late into the night after everyone has gone to bed...:bigeyes:
OH GOD, YOU'RE RIGHT! I think its happening and I haven't even finished the amp yet!:sigh:
*Grabs closest guitar and tries to barre Bbmaj7. Can only manage to conjure farty noises from the strings* Crap!:ill:

I think there's a country song waiting to be found somewhere in that narrative...
And we could write it! ... Except that we like the key of F for country and barring any form of Bb for the bridge or in a walking line is a non starter. :xeye: Lol :rofl:
 
...Keysight (nee Aglient nee HP) 1000 X-series CRO
Thanks for the link!

These 'scopes seem very fast and have been taught to do some clever tricks. Eight bits (vertical) is better than seven, but as the attached image shows (taken from page 6 of the above link), the nasty jaggies are still very much in evidence, most particularly in the pink (summed) trace. :(

Personally I'd rather have 10 bit vertical resolution (1024 steps) and only 100 kHz bandwidth, but the entire industry focus is in the opposite direction: more bandwidth, more speed, and who the heck needs better vertical resolution in a world where most circuitry is digital?


-Gnobuddy
 

Attachments

  • Keysight_Screen_Jaggies.png
    Keysight_Screen_Jaggies.png
    679.6 KB · Views: 51
I don't think I'll need something capable of "digital tricks" for my new hobby just yet, but on that note...
I was vague earlier, but there really are some useful digital tricks. One of them is measuring and displaying peak-to-peak voltage at the push of a button (you don't have to count graticule divisions, multiply by volts/div, carry the 2, forget the 5, and wonder how you came up with 39 million volts. :confused: )

Another great trick is that my Rigol has four channels, and each of the four traces on screen has a different colour, so you can instantly tell what's what. Imagine simultaneously monitoring power supply voltage, screen grid voltage, input audio on the control grid, and output signal on the anode. You can really see "sag", output clipping, what the screen grid resistor is doing, etc.

There are other good tricks too, some which I haven't used yet. Like a one-button FFT (Fast Fourier Transform), which shows you harmonics in the signal as you overdrive a valve. And since it's digital, it's also a storage oscilloscope - you can hit the amp with a single guitar note, and capture and freeze the amps response at up to four different circuit nodes to study at your leisure.

The lovely Tektronix analog 'scopes I used throughout my college years had no onscreen jaggies and were much easier on the eyes, but they had exactly zero of these features. Well, there were analogue storage 'scopes, but I didn't get to use one very often. (We had a Polaroid camera with a special adapter to mount over the 'scope screen, so you could photograph the signal for posterity, or at least, for you lab notebook.)
OH GOD, YOU'RE RIGHT! I think its happening and I haven't even finished the amp yet!:sigh:
See what I mean? You have only one hope, quit cold turkey. And mail me your amp immediately, so you're not tempted to re-start your new addiction. :cool:

...barring any form of Bb for the bridge or in a walking line is a non starter. :xeye: Lol :rofl:
I played guitar on and off for twenty five years without ever using a capo. Then I plucked up the courage to start going to my first jams. At the acoustic jams, everyone had a capo. The answer to the Bb conundrum is, capo at the first fret, and play everything in the key of E!

When I left the USA I thought I'd left country music behind for good. No such luck - it turns out several of the Canadian provinces love country music, too. And "cowboy music", whatever the heck that might be. :eek:

I know players who capo 4 and play in C major when they want to be playing in the key of E major - only because they cannot play a C#m, so capo 4 and play Amin is the solution. :)

The funny thing is, Nashville is full of virtuoso guitarists with incredible chops, mostly hired to play on 3-chord country songs. I really can't figure that out.


-Gnobuddy
 
As you get in to benchwork more, and aspects of performance such as noise floor, hum distortion, and frequency response curves from tone controls, you are likely to notice people using PC based audio software for scope, spectrum analyser, signal generation functions, and the benefit of using 24bit signal resolution. This seems to be finding traction for those who have an analog or modern digital scope, and start to realise what it can't 'see' or do with respect to some areas of interest.
 
...PC based audio software...
This is something I've been interested in for a while. About twenty years ago, I used PC-based commercial hardware / software tools such as the Audio Precision System One and MLSSA. Each one required an entire dedicated PC. So at the time that was very expensive equipment, but it was also eye-opening in it's capabilities.

Back then, I was using the AP System 1 and MLSSA with solid-state audio equipment. Signal voltages and impedances were low, and the tools came with hardware interfaces that could handle those signals

The situation seems very different when it comes to valve audio circuits. The enormous gulf between signal levels and input and output impedances of PC sound cards, and those in valve guitar amps, is one of the things that's kept me from getting started down this road.

For instance, it would be nice if the measurement tool (PC + software) put out a big enough AC signal to overdrive a single 6V6 output valve, say, maybe a 6L6. Say at least 50 volts peak-to-peak, and preferably up to 100 Vpp. But sound cards are limited to maybe a couple of volts, peak to peak, maximum output. (My audio signal generator doesn't do a whole lot better, maybe 20 volts peak to peak maximum.)

In the same way, the PC measurement system would have to cope with the possibility of a 300-volt (peak-to-peak) signal from a small-signal pentode, and, worst case, maybe 1000 Vpp from an output valve anode. We'd also want at least a one megohm input impedance, preferably ten times that. Meantime, the line-in jack on the PC wants 100 mV pp, and may only have 10k input impedance. And the user must be kept safe from the 1000 V at all costs.

Then there's the ground-loop issues. It seems to me that transformer coupling between PC sound card and valve amp test interface might be necessary to provide isolation, otherwise there will be ground loops, hum, and PC switching power supply noise to worry about.

So it seems to me that some sort of hardware interface between PC and valve guitar amp is virtually a necessity. And I'm a bit stumped on exactly how to put this together, and exactly what is required. Using real 'scope probes between valve amp and interface seems like a no-brainer, but there is a lot more to figure out, from scaling amplifiers and switchable attenuators to overload indicators and coupling transformers, not to mention some way of limiting the maximum signal that gets fed to the sound-card's line-in jack.

At about this point, I usually realize that designing and building one complicated electronic project (the interface) so that I can then design and build the electronic project I actually want (guitar amps), is likely to take more time, energy, and finances than I can muster. So I end up going back to the trusty old signal generator, scope, and pocket calculator.

So I have a question for those of you who are already using PC sound-card measurements with your valve audio circuits: how are you dealing with all these hurdles?


-Gnobuddy
 
I've only rarely needed more than the 2Vrms max output of my interface. Although I have used a cheap battery powered class d amp module with resistor loading to generate a much larger floating signal level, an analog amp would likely be needed for linking a soundcard to DUT.

I battery power my interface to remove one ingress point of mains earth loop noise. I use a battery powered USB isolator module (based on adum4160) when I need to isolate the PC to a really low level, although a laptop is the other solution.

There can also be a need to screen the interface from bench environment noise (I use 2 aluminium bbq baking trays, one for the top and the other for the tub, with holes punched through for USB and probe cables, and a grounding clip from probe BNC metal to trays, to provide the lowest repeatable noise floor setup - and the EMU0404 gets pretty low).

For most interfaces, the signal out and in are not isolated, so there is a need to try and identify a common DUT ground, and not just assume there is an internal connection in the amp - given that some valve amps have a floating speaker winding.

The cheap 100:1 probe I purchased didn't quite provide a flat response with my 1Meg interface input, but that was remedied by lowering the smd cap value shunting the compensation capacitor trim in the probe. As far as voltage capability, it's fine for any speaker output level and ground referenced nodes and even lower arms of a cathodyne, but similar to a normal scope scenario I wouldn't be connecting it to B+ or plate circuits without an additional coupling cap. And a 100:1 has negligible loading on pretty much any valve amp node.

You have to know how to use the PC operating system audio controls/drivers. My setup requires turning off any other audio device (speakers, microphone etc).

An interface that doesn't provide a high impedance input, or atten/gain controls, or a dc input, would be a pain for me.

I don't know of a cheap current model commercial USB interface, although there are a few forums that have threads on commercial interfaces that are highlighting good ones.

Although REW is aimed at speaker audio related assessment, it is effectively universal for audio work - with some forums enjoying the discovery process:
REW Update

There are a few analog buffer projects that act as the go-between from a probe to a standard soundcard - I used a 1998 vintage magazine buffer up until a few years ago.
https://dalmura.com.au/static/K2875%20Soundcard%20%20Preamp%20EA%20Aug%201998.pdf.
 
Last edited:
See what I mean? You have only one hope, quit cold turkey. And mail me your amp immediately, so you're not tempted to re-start your new addiction.
Nah, it's already too late for me, no sense in fighting it. Good try though lol :p

The answer to the Bb conundrum is, capo at the first fret, and play everything in the key of E!
I actually started my music life as a bass player, so the finger strength and dexterity to barre that first fret on the acoustic has never been a problem for me. I had much more trouble developing a lighter touch for barre chords when I started electric guitar. I would always fret with too much pressure on the lighter strings and pull everything sharp lol. I do like the capo for moving things to Emaj shapes on the acoustic occasionally though because I can let the high E and B strings drone on pretty much every chord. Low E too sometimes if it fits the song. Makes for a very nice full and open sound if I'm carrying rhythm by myself. :up:

The funny thing is, Nashville is full of virtuoso guitarists with incredible chops, mostly hired to play on 3-chord country songs. I really can't figure that out.
I share your lament on this. In general it seems there's not much popular music in any genre these days that strays from the I, V, vi, IV progression + a few variations of that. No doubt there are tons of wonderfully talented musicians and song writers out there, but you won't hear them playing much new and fresh on the radio anymore. You have to find other music outlets to hear people playing anything interesting these days. :soapbox:
 
On the subject of oscilloscopes; clearly trobbins and Gnobuddy are conversing at a level way beyond what I can follow without some more serious google research :spin: lol. However I appreciate the amount of info and experience being laid out here. I intend to finish my next round of mods on the TS and surrounding parts of the circuit before delving deeper into more equipment. Especially potentially expensive equipment :$::faint: . But when the time comes to buy one is there any merit to purchasing a cheaper digital scope ($100 or so) that can handle the basic functions while I learn how to use it. Then upgrade to something fancier later once I have a firmer grasp on correctly taking measurements and analyzing the data? Or is it necessary to spend $400 and up for something with more bells and whistles right away?
 
Good try though lol :p
:)
I actually started my music life as a bass player, so the finger strength and dexterity to barre that first fret on the acoustic has never been a problem for me.
When I first began playing guitar, teaching myself from library books I borrowed (no Internet or PC yet!), it took a while to build up enough left-hand strength to play barre chords. But everything else was hard too, at first. This was just one more hard thing that I knew I'd have to work at until it wasn't hard any more. So of course I did just that.

From the books I was reading, clearly this was just part of learning the craft, and everybody had to go through it. The alternative was to remain a novice permanently, and who would find that acceptable?

Well, it turns out, quite a lot of people do find it acceptable. I just didn't realize that until I began going to jams two and a half decades later. It baffles me to meet people who proudly tell me they've been playing guitar for fifty years, but they still believe barre chords are some exotic and difficult thing that is beyond their reach. I just don't get that at all. :confused:

It's the same story with music theory, or at least the rudiments of it. How can you play guitar for 50 years and not even get curious why "key of A major" means "Your country song will have the chords E7, A, and D in it"?

It's all very confusing to me, particularly because some of these people clearly do have musical talent and a good ear. Why limit yourself so badly, for so long? How do you not even realize you've limited yourself?
I had much more trouble developing a lighter touch for barre chords when I started electric guitar. I would always fret with too much pressure on the lighter strings and pull everything sharp lol.
I had that problem too. Eventually I re-strung my main acoustic guitar with a 0.010 gauge set of strings, and re-learned my right hand technique to get good tone with the thinner strings. Because of the longer scale length, the acoustic string tension is still higher than a Les Paul with an 0.010 set of strings, but the difference isn't as huge as before.

You have to find other music outlets to hear people playing anything interesting these days. :soapbox:
I used to search through NPR's "Tiny Desk Concerts" for musical acts with some talent. I used to find them every now and then (like the wonderful group Lake Street Dive), but lately even the Tiny Desk concerts are mostly a stream of dreary dreck.

Oh, well!


-Gnobuddy
 

PRR

Member
Joined 2003
Paid Member
You can learn a lot from a 'scope MUCH cruder than any you can buy today.

My Eico 427 claimed 500kc but wouldn't sync 50kc. I then had a fine Leader 5MHz, but supplemented it with a basket-case 'scope made in 1937, gutted before I got it, and rebuilt with $1 parts and almost able to reach 20,000Hz. You can see the shapes of the waves, clipping, bentness. While fancy 'scopes also extract numbers from waves, I can get OK numbers with a VTVM.
 
Sorry about the detour into PC measurement talk. :eek:

But for the purposes of this thread, I would say the take-away is the same thing I said before: while using a PC sound card as a 'scope for your amp project seems tempting and easy, in practice it is complex and difficult. For now, I'd stay away.
...when the time comes to buy one is there any merit to purchasing a cheaper digital scope ($100 or so)
As long as it's a real oscilloscope, just about any model should be able to do what you want.

By "real oscilloscope", I mean a large box with actual knobs and switches and BNC connectors on the front panel, which accept actual oscilloscope probes.

It can be an old analogue Tektronics, Phillips, Hitachi, et cetera - all of them made good lab scopes that were used in universities and industry about thirty or forty years ago.

Vacuum tubes are now well over a century old, so forty year old 'scope technology is not just cutting edge for tube work - it's beyond cutting edge, outright science-fiction fantasy from the far future. And you're not even dealing with exotic tube stuff (radio, microwave, etc). You are working with audio signals (much lower frequency than radio and TV), and that too, guitar signals, which are even lower frequency than "proper" audio. This was "easy" for a good 'scope even fifty years ago.

What I would say is, avoid the toy "oscilloscopes" that are smaller than a cellphone and have a tiny LCD screen. This one, for example: Mini 2.8 Inch TFT LCD Screen Display Digital Storage Oscilloscope DS211(Color:black): Amazon.ca: Gateway

In my opinion, those little toy 'scopes are utterly useless for working with valve circuitry (and may be utterly useless, period, for any actual use other than play.) The circuits in your guitar amp are high voltage, and high impedance; you need a 'scope that, above all, keeps you safe. The toy scopes can't even do that, never mind making usefully accurate measurements.

Here is a more professional looking oscilloscope - but also useless, even dangerous, for your purposes: Handheld Oscilloscope 2CH 40MHz 250MSa/s Type C & Multimeter 2in1 Hantek 2C42 | eBay

The giveaway is the type of input connector: this thing seems to use banana plugs, like a DMM. That can be deadly if you're measuring a 200 volt valve anode, and the banana plug pulls out of the little 'scope. Proper BNC connectors twist and lock in place - they won't come out from a tug on the wire.

IF it's in good condition, something like this old Tektronix is excellent for your purposes: #14 Tektronix Model 2246 MODA, 4 Channel Oscilloscope. Estate Sale. | eBay

(But the Tektronix will be heavy, and takes up quite a lot of room. Not ideal for our contemporary cramped living conditions.)

I just stumbled across this little gem: CircuitGear Mini – Syscomp Electronic Design

The price is right, it accepts real 'scope probes, it has 11 bits of vertical resolution, and the software is open-source. Those are all good things. But I have zero experience with it, so I really can't offer any advice. (Also, you need a PC or laptop to use with it, since it has no display of it's own.)

I haven't kept up with the rapidly changing world of current digital scope offerings, but I think I know where you can find some people who have. Are you familar with the website "EEV Blog"? If not,that is a great place to find information on current 'scopes, and get your questions answered: EEVblog – No Script, No Fear, All Opinion

If you ask for advice there, just make it very clear that you are only interested in using the 'scope for audio signals, and don't need a bandwidth bigger than 100 kHz.


-Gnobuddy
 
Last edited:
...accept actual oscilloscope probes.
I should have explained that using real oscilloscope probes is very important for several reasons. Firstly, and this is literally of life-and-death importance, they come with insulation designed to protect you from high voltage (exactly how high, will be in the specifications.)

Secondly, they are designed to accurately and safely transfer the signals from your valve amp into the delicate input of the oscilloscope. This means (a) not loading down the circuit you're trying to measure, (b) dividing down the big voltage if necessary, and (c) working accurately across a wide range of frequencies, so that the waveform of the signal is preserved, and you can make reasonably accurate measurements of signal strength, etc, at any frequency within your oscilloscope's capabilities.

You can buy scope probes that divide the signal voltage by 10, divide by 100, or don't divide at all. To work safely with the rather high voltages in a valve guitar amp, you probably need divide by 100 probes; confusingly, these are usually known as "x100" probes, because you have to take the voltage displayed on the 'scope and multiply by 100 to calculate the actual voltage going into the probe.

'Scope probes look like an ordinary ball-point pen with a wire trailing out of one end, but they are more sophisticated pieces of engineering than they appear to be. I've seen DIY high-voltage probes on the 'Web: a very bad idea! You don't want to find out the hard way that the insulation in your home-made probe fails when 300 volts is applied to the tip.

Bottom line: for use with valve circuitry, don't buy any kind of 'scope unless it uses proper oscilloscope probes!

Some 'scopes come with probes. In other cases, you can buy aftermarket 'scope probes from electronics suppliers like Digikey. You will probably save money if the 'scope comes with probes. But you may need to shell out for at least one x100 probe, which is a sufficiently rare beast that you are unlikely to get one with your 'scope.


-Gnobuddy
 
Sorry about the detour into PC measurement talk.
No apology necessary at all. This thread has veered off down all sorts of friendly non related avenues whilst my amp project lingers in waiting lol. It's all good :D

I used to search through NPR's "Tiny Desk Concerts" for musical acts with some talent. I used to find them every now and then (like the wonderful group Lake Street Dive), but lately even the Tiny Desk concerts are mostly a stream of dreary dreck.
Used to love Tine Desk, they've had some good ones on there for sure! Lake Street Dive has a fun R&B revival kind of sound that I dig, I've enjoyed their music before for sure. If you like that sound then St. Paul & The Broken Bones, The Alabama Shakes, Nathaniel Rateliff & The Night Sweats, Leon Bridges, Hozier, and Chris Stapleton all play in that similar vein of music as well. I'm pretty sure some of them have been on Tiny Desk before but I can't remember which ones...
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.