I finally decided to pull out the scope I bought last summer and put it to use. Nothing special but 15mhz dual-trace is more than enough for stompbox troubleshooting. I won't profess to having any skill in using them, but I figure if I'm seeing an LFO waveform on the screen, I'm doing something right. I wouldn't trust my judgment of the signal amplitude, but at least I'm seeing something. Always interesting to see the waveform of fuzzes and octave-up units as the signal dies away (one of the probes I have for the scope ends in a moulded 1/4" phone plug so I can just plug pedals in directly).
Currently, I'm looking at the LFO output of modulation effects, and specifically the one-op-amp LFO found in the Phase 90. In one case, I get something that looks almost like a sine wave, but has a bit of a spike at the bottom of the waveform. In another, I get something that looks more like an attack/release transient generator than an LFO. It rises exponentially, but then has a "cliff" and falls exponentially, such that it looks a bit like an ocean wave sweeping to shore towards the right. So neither of the circuits I looked at yielded true sine or triangle waves. The circuits sound almost right, but not perfect.
Two questions:
1) Is the P90 LFO supposed to generate a true sine or triangle output?
2) If it is, are there particular components to tinker with to improve the waveform shape?
3) Is what I see potentially a function of tapping the LFO signal at the wrong point on the board, or setting up the scope wrong?
I have a couple of Ross phaser boards that use the LM13600 for LFO that I'm not entirely pleased with the feel of the sweep, so they will be next on the block once the P90 LFOs are optimized.
Does your scope probe have a compensation trimmer built into it?
https://www.quora.com/What-is-oscilloscope-probe-compensation-What-is-it-needed-for
The ocean wave shape is what you get from a P90 style LFO rather than a true triangle, so it looks like you're probing in the right place.
Going by this schematic http://www.home-wrecker.com/ross_phaser_orange2.png (http://www.home-wrecker.com/ross_phaser_orange2.png) you want to probe across C16. The output of the LFO is the shape made by charging and discharging the cap through R37 and the rate pot, so you get an exponential rise and fall. It's basically like alternating between these two circuits http://www.electronics-tutorials.ws/rc/rc_1.html (http://www.electronics-tutorials.ws/rc/rc_1.html), http://www.electronics-tutorials.ws/rc/rc_2.html (http://www.electronics-tutorials.ws/rc/rc_2.html)
Quote from: Mark Hammer on September 06, 2017, 01:05:22 PM
2) If it is, are there particular components to tinker with to improve the waveform shape?
3) Is what I see potentially a function of tapping the LFO signal at the wrong point on the board, or setting up the scope wrong?
Mark, some factory schematics for modulation pedals had info on tesing and adjusting waveforms. You might try looking at some of them. I think it was mostly Boss that had that info.
Scope channel has to be DC coupled to see an accurate low frequency slope. If AC coupled, it's a high pass filter and distorts non-sine waves. It can be useful to see what your LFO looks like with AC coupling, so you can leave the scope on AC for general low level signal hunting where the DC level would otherwise throw your trace off the screen.
Dual amp LFO's where one is Schmitt trigger and the other an integrator (cap in feedback) produce a proper buffered triangle from the integrator, although you can get a spike at the triangle peaks caused by the Schmitt trigger switching transient. In the single op-amp type, you only have the Schmitt trigger using the amp and the integrator is being done directly by a cap, hence the exponential sloped wave.
A good Sine is not easy.
Whatever, the wave should be "smooth".
I feel Tremolo should be Sine-like; it should linger near the extremes.
In a Phaser, the real effect happens when "moving". A triangle may be appropriate.
The exponential triangle may have some pleasure on the ear (up and down are not the same) and may save a part.
Sine oscillators of low cost will often control amplitude by clipping, often on one side. This can give a glitch. If the glitch is in the low-gain side of a tremolo, it may not be noticed.
> I wouldn't trust my judgment of the signal amplitude
Keep a 1.5V battery handy. Remember where the DC setting is on the 'scope. Read the battery. The jump is 1.5V. Most guitar will be under 1.5V peak to peak. Many LFOs will be several times bigger. A 9V battery may be your ruler for bigger waves.
Quote from: anotherjim on September 06, 2017, 02:35:04 PM
Scope channel has to be DC coupled to see an accurate low frequency slope. If AC coupled, it's a high pass filter and distorts non-sine waves. It can be useful to see what your LFO looks like with AC coupling, so you can leave the scope on AC for general low level signal hunting where the DC level would otherwise throw your trace off the screen.
Dual amp LFO's where one is Schmitt trigger and the other an integrator (cap in feedback) produce a proper buffered triangle from the integrator, although you can get a spike at the triangle peaks caused by the Schmitt trigger switching transient. In the single op-amp type, you only have the Schmitt trigger using the amp and the integrator is being done directly by a cap, hence the exponential sloped wave.
<insert epiphany here> I had it on AC, assuming that since it was a changing voltage, that AC would be appropriate. I'll try it on DC when I get home from work. Thanks.
> assuming that since it was a changing voltage, that AC would be appropriate.
Where is the line in the sand between AC and DC?
There isn't any. Geo-scientists treat 0.1Hz signals as AC. Wall-supply "DC" power has varying components past 600Hz.
Most 'scopes on AC mode start to fall below 20Hz. Most trem-things run nearer 10Hz. You are on the edge. Using "DC" will preserve the slowest changes better.
If there was a tiny AC on a big DC, like the signal at V1a of a guitar amp, then you might wish to discount the big DC so you could see the little AC better. But in an LFO the signal is typically a very large fraction of the supply voltage, so reading AC and DC together (on the "DC" mode) is no problem.
Good point. DC setting for LFOs. Otherwise the scope can lead you down a very strange path.
Took a bit of twiddling, and setting the LFO rate and scope sweep juuuuuuuussst right, but setting it to DC, instead of AC brought up a nice triangle, just like Jim said. And setting it back to AC yielded the "ocean wave".
Ya learn something new every day. Thanks, gents.
Mark, you know my life is incomplete until you tell me whether you compensated your probes. :icon_wink:
How would I have compensated them? FWIW, the scope shows "1M/35pf" on the inputs. I'm not well-versed enough in such matters to know what the implications of that are.
Normally there is a square wave test point out on the scope and a trim pot on the probe somewhere. You tweak it for the squarest looking result, the idea being to flatten the high frequency response.
Quote from: Mark Hammer on September 06, 2017, 08:55:59 PM
How would I have compensated them? FWIW, the scope shows "1M/35pf" on the inputs. I'm not well-versed enough in such matters to know what the implications of that are.
Maybe this will help a little. Your 1M/35pF is on the right hand side of this figure. In simplest terms, you adjust c_comp (with a screwdriver) until a known square wave at the tip of your probe looks like a square wave on your screen.
(http://static.tiepie.com/gfx/Classroom/basics/ProbeAttenuating.png)
Edit: Yeah, what Larry already said. :icon_wink:
Nothing to adjust on either of my probes. The scopeis a Philips PM-3207. There is what appears to be a contact point just below the power switch that is labelled "probe adj". Is that what you are referring to?
What range of capacitance what the adjusting part be?
(http://www.philipsradios.nl/forum/images/uploaded/201305191741315198f2ab6530b.jpg)
"probe adj" is likely your scope's built-in test signal. Go ahead and connect your probe tip to it and see if you have a nice square wave.
Nuthin that I can see. May need some twiddling, and it's getting kinda late. I'll check tomorrow.
> Good point. AC setting for LFOs.
"DC" gets my vote.
> compensated your probes.
Irrelevant where Mark is sitting. 1X probes don't compensate. 10V probe compensation affects the response in and above the top of the audio band, NOT down in the LFO zone where Mark is today. Even his glitching is mostly below the compensation point. And he may be using the 1X probe, no compensation. Or some more modern scheme, perhaps with buffer in the probe.
(https://s26.postimg.org/z2jpepgfp/PM3207-1979.gif) (https://postimg.org/image/z2jpepgfp/)
Typo! DC setting for LFOs. I remember it well - whacked and weak waveform, yet I gave up on the waveform and finished the circuit and it sounded pretty good!
Deciding whether something is "AC" or "DC" on a scope is tough. You are, in effect, trying to guess what the designer of the input preamplifier for the scope was thinking, and whether his boss was standing over his shoulder, and about what. The cap that blocks DC on a scope preamp has to be of high quality, and is expensive, at least the way HP and Tek used to do it. When in doubt, use DC.
There's a subtle problem lurking in here. LFOs don't exist in a vacuum. The circuit that is being modulated has some transfer function of volts of LFO to [units of something] of effect. That's not necessarily linear. It too has quirks of not exactly following the exact voltage of the LFO or having the same audible change in (whatever) per small change in LFO voltage at all points on the modulation.
And it can "saturate", not following the LFO waveform if it's over X or less than Y. There's a lot of fussiness in analog synth circuits to make sure that voltage controlled things are either accurately linear in following a control voltage, or accurately exponential in frequency-control situations.
Unless it's in a well-designed synth module, you can't depend on the audible effect accurately following the waveform of an LFO.
LFO circuits don't exist by themselves. They have to take into account the needs of what they control, and (for instance) whether there is a fixed DC level needed by the Controll-ed One, or if the LFO must expand only UP from some floor DC level, or only down from some ceiling DC level. Tremolo circuits show all three of these quirks, and you can sometimes tell whether a trem does tremolo up, tremolo down, or tremo up and down by listening. It gets hairy if you're driving a variable-C or variable-L resonance, where you control the frequency linearly by the square of the component value, and when what you want is neither linear nor square-law, but exponential.
One or two x10 probes ought to be on your shopping list Mark. Those are the ones that need calibrating which will need doing when checking BBD clocks and other fast stuff. I have switchable x1/x10 probes which add another chance for error if the switch gets accidentally knocked.
As even the cheapest DMM might have 10M loading, and we can see even that causing a lower voltage reading on a 1M op-amp input, your 1M scope input is going to halve the voltage and seriously offset the op-amp * , since the scope 1M is to ground not the op-amp Vref. x10 probes up the load on the circuit to 10M, but you and/or the scope need to know that the voltage reading will also be x10 lower than with the x1 probe.
* Caveat, the scope input may not be grounded (is floating) to AC supply safety ground and the circuit under test may not be either. Where the scope AND circuit are grounded, I find it better to not connect the probe ground to the circuit ground or else you have a ground loop which can inject all sorts of noise on the scope trace.
If either scope or circuit are not grounded, then a ground needs connecting between them. It is actually better to use a separate wire - a banana jack or binding post on the scope to a croc' clip or use the screen of an unused BNC connection. That leaves your test probe free to move around more and reduces ground return effects in the probe.
Quote from: PRR on September 06, 2017, 09:58:34 PM
(https://s26.postimg.org/z2jpepgfp/PM3207-1979.gif) (https://postimg.org/image/z2jpepgfp/)
Interesting. Thanks for that, Paul. I'm also glad I paid twenty 2016 dollars for it, instead of seven hundred and ninety-five 1979 dollars for it.
I guess with the money saved, I'll pop by Active Electronics this weekend and see what they have in the way of probes. :icon_smile:
What he said.
I got into the habit of using two scope probes, even if only one was needed, so that one of them carried ground and I could use the other for probing voltages. This introduces noise for situations where the ground is noisy and the ground point is "far away" from the actual signal being tested. You see signal plus noise, which is what the probe is actually seeing.
This won't be a problem on pedals, likely. For low noise measurements, especially of fast signals, the probe really needs to be two-headed, with the probe ground going to the nearest ground point to what's being measured.
I have fond memories of connecting the scope ground to the Wrong Thing while working on a switching power supply design and watching in horror as the insulation melted off the wire on the ground probe. Seems that the scope was carefully third-wire grounded, the probe ground was tied to that, and the point I was "grounding" with the scope probe was also tied to the AC line - but not to AC sefety ground. Ooooops... :icon_lol:
Just saw your post as I hit "submit". Marlin P Jones sells decent(ish) 10:1 probes for ~~US$20, and I see them at various replacement parts places for a similar price. Tek used to nick you for $100 in 1970s dollars for similar stuff.
Yep been there Jim.
Excellant thread. :)
Is something like this suitable? https://www.banggood.com/DANIU-P6100-Oscilloscope-100MHz-PKCATI-BNC-Clip-Probes-Clip-Cable-p-1157611.html?rmmds=category
"Banggood"? When the sniggers abate, look very much like the probes I bought from Rapid. Generic Chinee OEM I expect.
I've had other cheap ones where the probe spring cap flies off at every opportunity or fails to make contact with the actual tip underneath, so I don't know how those ones are for quality.
Note the little coloured ID rings you get with them (yellow shown on that catalogue page) - very useful with more than one scope channel on the go.
Also note the voltage rating which is typical. Only x10 is up to poking around a tube amp B+, that is if the scope input is good for it too.
Also keep in mind that, with Banggood, you are ordering directly from China. I recently ordered a laser engraver through them, which took about 3 weeks to arrive. Also, occasionally, imported goods get mysteriously stuck in Customs or require that you pay an import duty to receive them.
Mark, that 'scope seems to be "auto range". This does make reading actual voltage swing trickier, because the too-smart beast is changing the gain "for you". I did not see an Operator's Manual. I dunno if there's LEDs to tell you what gain it picked. A hasty peek at the input stage crossed my eyes, and I would not be *sure* that ordinary probes will work right. If you got the manual with it, take it to the throne room for a good read.
_____________
> ordering directly from China
His link suggests he gets it from a Canadian warehouse.
However it is also quoting 7-20 days, which does smell a bit like Air China over the pole and into the Customs pit.
Quote from: PRR on September 07, 2017, 02:04:39 PM
His link suggests he gets it from a Canadian warehouse.
"CN Direct" means it is shipped from Banggood's warehouse in China.
I've ordered from them before. It was fine. Just slow. I gather it came on the proverbial boat.
Quote from: PRR on September 07, 2017, 02:04:39 PM
Mark, that 'scope seems to be "auto range". This does make reading actual voltage swing trickier, because the too-smart beast is changing the gain "for you".
Looks to me like the automatic level is for *triggering*, not the display. The picture Mark posted shows fairly typical V/div settings for both channels.
Yeh, Auto trigger means it's always sweeping the X axis rather than waiting for an edge trigger input (+ or - going from A or B).
This is when you always get a display but waveforms have unstable position since they are rarely in synch with the timebase sweep.
Normal single channel probing use that channel as trigger source with +trigger for a stable display, but only with signal present.
If tracing through an audio path, use one channel as trigger source connected to circuit input (chB usually) and the other to trace through the following stages. Display both channels (A+B) to see differences from input wave including phase shift/inversion.
ADD mode is the sum of A & B voltage in one trace (check for balance between differential outputs, they should perfectly cancel if exact 180deg phase and equal amplitude) . Invert chB with ADD and you have a difference amp input. Diff amp is excellent when you cannot use a common ground between scope and circuit as both probe tips only are used to provide ground isolated + and - signal connections. With x10 probes, this is one of the safest ways to work (safe for the equipment, not you).
X Via A on the timebase switch removes any horizontal sweep and chA voltage determines horizontal position. With signal on both channels, the CRO can draw shapes according to the differences between the 2 signals...
https://en.wikipedia.org/wiki/Lissajous_curve.