100k input resistance... isn't

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!

Moderators: Kent, Joe., luketeaford, lisa

User avatar
neil.johnson
Super Deluxe Wiggler
Posts: 1519
Joined: Thu Apr 08, 2010 3:51 pm
Location: Cambridge, UK
Contact:

100k input resistance... isn't

Post by neil.johnson » Sat Nov 21, 2020 3:11 pm

I keep seeing so many posts where folks say such things as:
A typical Eurorack input has an impedance of 100k.
The more I see this, the more I realise this is such a dangerous assertion. Very rarely is this assertion actually correct, and so any discussions based on it are likely wrong from the outset.

It's not even worth pointing out any one example of where this is not true -- there are too many! It's not about shaming or exposing, it's about poor design, especially in places where it matters: CV inputs.

I would be curious to understand why, when it is relatively straightforward to buffer an input before using it, that many designs simply don't? Is it cost? Is it lack of understanding? Is it not as important as I think it is??? Curious to understand...

Neil
Random ramblings :: http://www.njohnson.co.uk
SSI parts and more :: https://www.cesyg.co.uk

User avatar
cackland
Super Deluxe Wiggler
Posts: 2258
Joined: Thu Dec 28, 2017 5:42 pm
Location: Los Angeles, California

Re: 100k input resistance... isn't

Post by cackland » Sat Nov 21, 2020 4:40 pm

Can you provide a reasoning behind why the 100k input resistance may be incorrect and what might be a better input impedance?

socom93
Common Wiggler
Posts: 79
Joined: Sun Mar 24, 2019 1:19 pm
Location: France

Re: 100k input resistance... isn't

Post by socom93 » Sat Nov 21, 2020 4:57 pm

neil.johnson wrote:
Sat Nov 21, 2020 3:11 pm
I keep seeing so many posts where folks say such things as:
A typical Eurorack input has an impedance of 100k.
I would be curious to understand why, when it is relatively straightforward to buffer an input before using it, that many designs simply don't? Is it cost? Is it lack of understanding? Is it not as important as I think it is??? Curious to understand...
Well without telling in which case you encountered such unbuffered input it is difficult to tell if buffering was needed or not, if it was laisyness or lack of knowledge.

As a designer, i can tell you 2 cases where you can do such things :
- In case you really don't have any space for another opamp ( but like really really, like even with every resistors japanese style)
- If you design it for yourself knowing what you will put in the unbuffered input.

User avatar
neil.johnson
Super Deluxe Wiggler
Posts: 1519
Joined: Thu Apr 08, 2010 3:51 pm
Location: Cambridge, UK
Contact:

Re: 100k input resistance... isn't

Post by neil.johnson » Sat Nov 21, 2020 5:19 pm

cackland wrote:
Sat Nov 21, 2020 4:40 pm
Can you provide a reasoning behind why the 100k input resistance may be incorrect and what might be a better input impedance?
For example, a variable CV input going to a 100k pot, with the wiper connected to a 100k resistor to an inverting op-amp. That input will only be 100k when the wiper is at the ground end of the pot, reducing to 50k with the wiper at the top (input) end. Obviously not for pitch, but could be for linear FM for example. I have seen some designs where a higher wiper load is used, e.g., 470k. It's an improvement - the input resistance drops down to 82.5k.

If a CV source with 1k output resistor goes to a passive mult and then to a VCO CV input and an adjustable input, once you get it tuned up it will then go out of tune as the variable input is changed.

Neil
Random ramblings :: http://www.njohnson.co.uk
SSI parts and more :: https://www.cesyg.co.uk

User avatar
neil.johnson
Super Deluxe Wiggler
Posts: 1519
Joined: Thu Apr 08, 2010 3:51 pm
Location: Cambridge, UK
Contact:

Re: 100k input resistance... isn't

Post by neil.johnson » Sat Nov 21, 2020 5:23 pm

I dunno, maybe I'm over-thinking this. Maybe it's just funny: the "100k input impedance" mantra, and oh look a 100k pot, but then that's all blown up by the wiper resistor. When a simple opamp buffer would (a) guarantee a constant 100k input impedance, and (b) allow the use of lower valued pots, e.g., 10k, so that decent wiper loads can be used (I generally aim for a wiper load 10x that of the track resistance), which also helps to preserve the track law.

Neil
Random ramblings :: http://www.njohnson.co.uk
SSI parts and more :: https://www.cesyg.co.uk

User avatar
MikeDB
Wiggling with Experience
Posts: 326
Joined: Fri Feb 08, 2019 8:28 am
Location: UK

Re: 100k input resistance... isn't

Post by MikeDB » Sat Nov 21, 2020 6:09 pm

neil.johnson wrote:
Sat Nov 21, 2020 3:11 pm
I keep seeing so many posts where folks say such things as:
A typical Eurorack input has an impedance of 100k.
The more I see this, the more I realise this is such a dangerous assertion. Very rarely is this assertion actually correct, and so any discussions based on it are likely wrong from the outset.

It's not even worth pointing out any one example of where this is not true -- there are too many! It's not about shaming or exposing, it's about poor design, especially in places where it matters: CV inputs.

I would be curious to understand why, when it is relatively straightforward to buffer an input before using it, that many designs simply don't? Is it cost? Is it lack of understanding? Is it not as important as I think it is??? Curious to understand...

Neil
I totally agree with you. Stating something is X volts without any idea of it's output impedance (rarely zero) or the impedance it feeds into is just a nonsense nowadays. I can understand it back in the 60s when RM was designing the first modulars when op-amps weren't affordable and even transistors had to be limited, but now you can even get a TL074 for 10p from RS so there really is no excuse not to buffer everything.
Personally I hate cats ! Charlie the jackdaw on my shoulder also hates cats as his mother was killed by one so I had to raise him until he set off to procreate his lineage.

User avatar
neil.johnson
Super Deluxe Wiggler
Posts: 1519
Joined: Thu Apr 08, 2010 3:51 pm
Location: Cambridge, UK
Contact:

Re: 100k input resistance... isn't

Post by neil.johnson » Sat Nov 21, 2020 6:38 pm

Another thread provides some examples: viewtopic.php?f=17&t=240534#p3394348

Both of those circuits vary their input resistance as the pots are swept from one end to the other.

Neil
Random ramblings :: http://www.njohnson.co.uk
SSI parts and more :: https://www.cesyg.co.uk

User avatar
KSS
Super Deluxe Wiggler
Posts: 4174
Joined: Mon Jan 25, 2016 7:28 am

Re: 100k input resistance... isn't

Post by KSS » Sat Nov 21, 2020 6:53 pm

This post is wrong on so many levels, with that said, let's get to it.
neil.johnson wrote:
Sat Nov 21, 2020 3:11 pm
I keep seeing so many posts where folks say such things as:
A typical Eurorack input has an impedance of 100k.
What's wrong wth this? It's more true than not. Your example below is for inputs with typical attenuating pot architecture. But are those the most typical? I think not, though it would fun for someone -not me!- to go to modular grid and count them vs what I think will prove to be for more common. Bare inputs leading directly to a 100K input resistor.
The more I see this, the more I realise this is such a dangerous assertion.
Dangerous? LMAO
Very rarely is this assertion actually correct,
see the more likely truth above. I believe it's your intial hypothesis that's off.
and so any discussions based on it are likely wrong from the outset.
Again, there are PLENTY of examples of true 100K inputs in all formats of the synth kingdom
It's not even worth pointing out any one example of where this is not true -- there are too many!
Something we can agree on. there *are* many examples of module inpuits which are *not* 100K . But that's a FAR cry from your ignoring all of those which ARE.
It's not about shaming or exposing,
Good. Because there's nothing to "shame" or "expose"<-- Well, I do think it's good you pointing out that there are differences. But to put such a bias on it works against your goal. It's too easy to shoot down the faults in your position, and unfortunately that also keeps the good part of your message from getting the exposure it needs and deserves. More on that later.
it's about poor design, especially in places where it matters: CV inputs.
No, it's not. And no, those aren't "special places where it matters". If you'd said pitch CV inputs then maybe I'd back off<-that part. Using appropriate circuits in both cost *and* function does not mean one always chooses the electrically 'best' design. Best depends on a matrix of options and results. Any good designer-engineer knows that to their core.
I would be curious to understand why, when it is relatively straightforward to buffer an input before using it, that many designs simply don't?
Are you? really? curious? You know damn well why. At least I hope you do!
Is it cost?
Emphatically YES. Why on earth should I or anybody else buffer every input? Next you'll be tellng us every synth needs to have balnaced inputs and outputs on every jack or audio connector.. :bang: Straightforward is not the only or best choice.
Is it lack of understanding?
Emphatically NO. It is well understood by those with even a meager electronics background. You know that.
Is it not as important as I think it is??? Curious to understand...
Bingo!. Ding! Ding! Ding! We have a winner. Chicken dinner!
We have half a century plus of proof that the people like yourself who keep bringing this up as some sort of pox on the synth universe <--It's NOT. are simply sharing an opinion of their position on the matter's value.

Haven't we learned from too, too many human arenas that 20% solutions satisfy 80% of customers? That better doesn't usually matter if the "problem" isn't perceived as such by the 80%?

We've also learned -and most of us watch out for- the first step in changing something is to elevate it -often artificially so- as a problem. Our gut check says, Hey, wait a minute. That doesn't really matter.

Like I said above, you have a good message and goal. Framing it as if we're all lost without embracing it is quickly seen for the falsity it is, and the meat and potatoes of your message is lost in the mix.<--Just like the results of all those "not really 100K's" is ALSO lost in the mix. Literally.
-----------------
In adapting the 1003 to my new format, I noticed that there was room on the bus IO -and panel- for the inverter which is otherwise buried in the output circuit. I decide to break it out to be used at will. When Dennis Colin designed it, he used a 10K input and feedback resistor for this typical OPA inverter. I changed those to 100K, knowing full well that DC had used 10K to reduce noise. Because when it was embedded, there was no need to condition for naked expectations. But had I not mentioned it here, no one, and I mean no one, would have ever noticed or questioned it. Because while the difference going from 10K to 100K for those two resistors *will* increase the noise, the end result increase will. not. ever. matter. That's not bad design. That's desiogning for the reality the module will face, and for the customer;s it's intended for, and at the price level they expect. <--Price here not really a thing. But still part of ANY good design. Oh, and that inverter input now IS 100K. ;) e pluribus unum. One of many.

User avatar
KSS
Super Deluxe Wiggler
Posts: 4174
Joined: Mon Jan 25, 2016 7:28 am

Re: 100k input resistance... isn't

Post by KSS » Sat Nov 21, 2020 7:15 pm

@Mike DB. Many here know the electrical reasoning, and how it makes sense.

But the fact an OPA costs little doesn't remove its current draw from the equation. Nor its footprint. Nor those of any and all associated components for every buffer.

And what do we gain in the end? A better system. More even results that are exactly as the engineer predicts.

SO WHAT. <--As in Big Deal. As in Why Bother?

This is like trying to design a guitar that never needs tuning.<-- I did once know a person who had made a mechanically auto-tuning guitar.
Plenty of people would benefit. But it's a solution in search of a problem.

Even for pitch. Anyone working with synths quickly realizes what they have to do to deal with the 'problems' of varying impedances. Of 1K output resistors. Even for pitch. Some will end up with a Hinton per channel fine trim. Another will do some effectively identical solution in their patching or workflow.

Just as they have already done -and just as we've all already heard- on thousands of albums and songs over four dozens of years.

Edit:
Musician, to Audio engineer: WTF? When I patch in a second VCO it all goes pitch wonky.
Audio Engineer. Yes. that's because there are resistors where they needn't shouldn't be. Incredibly poor design.
Musician: But aren't they all like this?
AE: Not all. But Yes, most are. More cargo cultism and bad design from people who don't know WTH they're doing!
M: Can it be fixed?
AE: Sure. Just needs a circuit mod and ultimately a redesign of the whole synth universe.
M: What's that gonna cost? For my part, right here? Not the whle universe thing.
AE: $$$
M: <-- This oculd go two ways. Depends on the musician -and their situation- what this line-answer will be. Some *will* go for it, others will plug onward as is.

But let's pretend we just told the musician instead it couldn't be fixed. What would they do?

The. Exact. Same. Thing. They've. Been. Doing. Already.

They'd deal with it. Discover and employ work arounds. Make great music for us all to listen to and enjoy.
-------------------------------------------------------------------------------------------------------------------------------

Again, it's not that the message isn't valid. It is. But it's already been 'solved'. If you try to say otherwise by pointing out posts where someone is complaining about it, I'd say they're either new to this. Which has been clear in many or most of the threads where this comes up.

OR, they're in the position of that last M: of that imagine conversation, and they are the ones who *will* go for it.

There are FAR fewer of those than the rest. Who just keep on keeping on making music.

User avatar
mrand
Common Wiggler
Posts: 219
Joined: Thu May 23, 2013 11:26 pm
Location: Whitehorse, Yukon
Contact:

Re: 100k input resistance... isn't

Post by mrand » Sat Nov 21, 2020 7:59 pm

Great topic, thanks for bringing it up.

I'm going to come right out and admit that I don't understand how to measure or calculate impedance, nor can i predict the effects of mismatched i/o impedances. In fact, sometimes I'm not even sure I know what impedance is, at all. For example I always wondered how a 100k input attenuator wired to a 100k input resistor could yield the professed 100k input impedance.

So there you have it. Add one amateur designer to your "lack of understanding" list!

But also, I would like to add that only rarely do i notice problems due to impedance missmatch among my modules. I've been spanked before, but even in that case, it was more of an annoyance than a deal-breaker.

Looking forward to any best-practice recommendations in this regard!
--ssdp--

User avatar
MikeDB
Wiggling with Experience
Posts: 326
Joined: Fri Feb 08, 2019 8:28 am
Location: UK

Re: 100k input resistance... isn't

Post by MikeDB » Sat Nov 21, 2020 8:53 pm

KSS wrote:
Sat Nov 21, 2020 7:15 pm
@Mike DB. Many here know the electrical reasoning, and how it makes sense.

But the fact an OPA costs little doesn't remove its current draw from the equation. Nor its footprint. Nor those of any and all associated components for every buffer.

And what do we gain in the end? A better system. More even results that are exactly as the engineer predicts.

SO WHAT. <--As in Big Deal. As in Why Bother?
More even results as you put it is one of the main reasons the Mini-Moog succeeded, as any top keyboardist could just look at it and know EXACTLY what sound it would make from any patch, whereas with a modular Moog, VCS3, ARP2600, etc it was hit and miss as to what sound you would get, or if you would get the same sound two nights running. Repeatability is something to bother about, especially if you are touring with a backup synth in case one fails. Rick Wakeman was touring with six of them at one point - imagine the nightmare for his engineer if they hadn't been totally interchangeable.

As for current draw, come on. Modern power supplies can source 10s of amps without complaint. Lots of people replace the TL072s in their mixers with 5532s - the only reason we didn't use them in the first place is the heat build up, not that the power supply couldn't keep up.

And problems with footprints disappeared years ago when we went surface mount. There might be a shortage of space for a TL074 and some 0402 resistors in an iPhone, but not in a Eurorack module.

Oh and trust me, if I'd have told an (unnamed) famous musician or his engineer something couldn't be fixed to their liking, they'd have thrown the whole mixer back at us and gone and bought another brand. Good design always wins.
Personally I hate cats ! Charlie the jackdaw on my shoulder also hates cats as his mother was killed by one so I had to raise him until he set off to procreate his lineage.

User avatar
KSS
Super Deluxe Wiggler
Posts: 4174
Joined: Mon Jan 25, 2016 7:28 am

Re: 100k input resistance... isn't

Post by KSS » Sat Nov 21, 2020 9:01 pm

All that said, Acts also toured with the other synths you listed.

As for the power and size, sure. but Neil's position was that ALL modules should have ALL IO buffered. And we still have *many* modules using TH, while the size pressure of Eurorack continues its -to me- silly march to micro unusability. and we hav increasing numbers of modules taking up more and more of that available power -along with muddying it for others- by the increase in uP heavy digital. They're both still generally valid statements to make in answer to a why isn't *everything* some other way.

User avatar
KSS
Super Deluxe Wiggler
Posts: 4174
Joined: Mon Jan 25, 2016 7:28 am

Re: 100k input resistance... isn't

Post by KSS » Sat Nov 21, 2020 9:12 pm

Mike DB wrote:Oh and trust me, if I'd have told an (unnamed) famous musician or his engineer something couldn't be fixed to their liking, they'd have thrown the whole mixer back at us and gone and bought another brand.
Sure. But that misses the point i was making. that all those non-famous, non money-is-no-object musicians quickly learn to deal with the vagaries and ideomatic symptoms they run into and deal with them. Lording over a big compay's AE is not an option for the,m. At least not an especially effective one.
Good design always wins.
If ONLY this WERE true. It's not. And we all know it. It *should* win. It *mostly* does. But just as often the better design does *not* win.

One only need think about this thread's existence to see the truth of that!

If good design had won, we wouldn't be having this discussion.. It didn't. Even over 50 years. It didn't. And so we debate.
--------------------
I will once again say I'm *not* disputing the electrical merits of the proposal. I *am* answering the 'why' of where we are, and the practical downsides to the proposed 'solution's acceptance n general practice.
Last edited by KSS on Sat Nov 21, 2020 9:14 pm, edited 1 time in total.

User avatar
MikeDB
Wiggling with Experience
Posts: 326
Joined: Fri Feb 08, 2019 8:28 am
Location: UK

Re: 100k input resistance... isn't

Post by MikeDB » Sat Nov 21, 2020 9:12 pm

KSS wrote:
Sat Nov 21, 2020 9:01 pm
All that said, Acts also toured with the other synths you listed.

As for the power and size, sure. but Neil's position was that ALL modules should have ALL IO buffered. And we still have *many* modules using TH, while the size pressure of Eurorack continues its -to me- silly march to micro unusability. and we hav increasing numbers of modules taking up more and more of that available power -along with muddying it for others- by the increase in uP heavy digital. They're both still generally valid statements to make in answer to a why isn't *everything* some other way.
Act did tour with those synths, and sound engineers tore their hair out every night trying to set the f***ing things up. Just listen to the various ELP videos on YouTube and you can easily tell when the Beast was having a good night or a bad one.

As for TH, obviously fine for prototyping or DIY (which I accept this forum is) but surely nobody considers a production run using TH anymore. Is it even possible to get TH boards loaded in quantity nowadays ? The last production TH board I designed was almost 25 years ago.
Personally I hate cats ! Charlie the jackdaw on my shoulder also hates cats as his mother was killed by one so I had to raise him until he set off to procreate his lineage.

User avatar
MikeDB
Wiggling with Experience
Posts: 326
Joined: Fri Feb 08, 2019 8:28 am
Location: UK

Re: 100k input resistance... isn't

Post by MikeDB » Sat Nov 21, 2020 9:16 pm

KSS wrote:
Sat Nov 21, 2020 9:12 pm
Mike DB wrote:Oh and trust me, if I'd have told an (unnamed) famous musician or his engineer something couldn't be fixed to their liking, they'd have thrown the whole mixer back at us and gone and bought another brand.
Sure. But that misses the point i was making. that all those non-famous, non money-is-no-object musicians quickly learn to deal with the vagaries and ideomatic symptoms they run into and deal with them. Lording over a big compay's AE is not an option for the,m. At least not an especially effective one.
And once those non-famous impoverished musicians break through into the big time, they ditch the crud and buy some better kit. Whilst Behringer have proven there's money to be made at the low end, most of the profits in the music industry are made at the high end. So what Neil and I are saying is try to think like a high end designer in the first place - it's worth your while.
Personally I hate cats ! Charlie the jackdaw on my shoulder also hates cats as his mother was killed by one so I had to raise him until he set off to procreate his lineage.

User avatar
KSS
Super Deluxe Wiggler
Posts: 4174
Joined: Mon Jan 25, 2016 7:28 am

Re: 100k input resistance... isn't

Post by KSS » Sat Nov 21, 2020 9:23 pm

Mike DB wrote:So what Neil and I are saying is try to think like a high end designer in the first place - it's worth your while.
I agree. I DID design for adding -optional- IO buffers on *every* IO of the 50SQ modules and system in my new format based on the 2500.

User avatar
neil.johnson
Super Deluxe Wiggler
Posts: 1519
Joined: Thu Apr 08, 2010 3:51 pm
Location: Cambridge, UK
Contact:

Re: 100k input resistance... isn't

Post by neil.johnson » Sun Nov 22, 2020 6:31 am

mrand wrote:
Sat Nov 21, 2020 7:59 pm
Great topic, thanks for bringing it up.

I'm going to come right out and admit that I don't understand how to measure or calculate impedance, nor can i predict the effects of mismatched i/o impedances. In fact, sometimes I'm not even sure I know what impedance is, at all. For example I always wondered how a 100k input attenuator wired to a 100k input resistor could yield the professed 100k input impedance.

So there you have it. Add one amateur designer to your "lack of understanding" list!

But also, I would like to add that only rarely do i notice problems due to impedance missmatch among my modules. I've been spanked before, but even in that case, it was more of an annoyance than a deal-breaker.

Looking forward to any best-practice recommendations in this regard!
All great questions!

Consider the CGS circuit from the thread above: viewtopic.php?f=17&t=240534#p3394348
There is a 100k input level pot, with 100k wiper load to a virtual earth.
When the wiper is at the 0% position the source only sees the track resistance, so 100k.
When the wiper is at the 100% position the source sees the track resistance in parallel with the wiper load, so 100k in parallel with 100k = 50k.
When the wiper is at the 50% midway position it gets a bit more complicated: split the track into two equal 50k halves, now the input sees the top 50k in series with (the bottom 50k in parallel with the wiper load = 33.3k). So that's 83.3k.

The Barton circuit is slightly different as the opamp gets involved.
When the wiper is at the 0% the source sees the 100k track resistance in parallel with the 200k input resistor = 66.7k.
When the wiper is at the 100% because of opamp behaviour (assuming it is operating in its linear region) then both sides of the 200k input resistor are at the same voltage, so no current flows through it, so it is to all intents an open circuit. So the load only sees the 100k track resistance.
When the wiper is at the 50% midway position you still have the 100k track, and because the other side of the 200k input resistor is at half the input voltage it looks like a 400k resistor, so 100k in parallel with 400k = 80k.

This issue is not about impedance mismatch (that's more of a concern at radio frequencies or in power circuits) but more about changing resistance - you build a patch, get it all in tune, sounds awesome :bananaguitar: - then you tweak an input attenuator and suddenly you're drifting out of tune :omg:

Hope that helps!

Neil
Random ramblings :: http://www.njohnson.co.uk
SSI parts and more :: https://www.cesyg.co.uk

User avatar
CLee
Super Deluxe Wiggler
Posts: 1754
Joined: Mon Apr 05, 2010 6:49 pm
Location: Chicago
Contact:

Re: 100k input resistance... isn't

Post by CLee » Sun Nov 22, 2020 9:49 am

Serge designed his modules with the protection/stability output resistors inside the feedback loop, so the varying loads are compensated for. Using banana jacks, he knew patches would stack and fan out outputs to multiple inputs. As the load increases on the output the op amp increases its gain, so as long as it doesn’t clip it stays in tune. The actual input impedance value doesn’t matter.

Too bad Euro didn’t do the same

User avatar
mskala
Super Deluxe Wiggler
Posts: 2717
Joined: Thu Mar 27, 2014 7:33 am
Location: Toronto
Contact:

Re: 100k input resistance... isn't

Post by mskala » Sun Nov 22, 2020 11:22 am

neil.johnson wrote:
Sat Nov 21, 2020 3:11 pm
I keep seeing so many posts where folks say such things as:
A typical Eurorack input has an impedance of 100k.
The more I see this, the more I realise this is such a dangerous assertion. Very rarely is this assertion actually correct, and so any discussions based on it are likely wrong from the outset.
This looks like an indirect way of calling me out for this thread: viewtopic.php?f=17&t=240512 in which I wrote those exact words. If so, it's pretty rude, and I'm not sure I should dignify it with a response. But it's true that other people have said similar things many times too.

"Typical" doesn't mean that it's going to be exactly 100k every time, as anyone who reads data sheets knows very well. Many Eurorack inputs, such as those using the very common pattern of an inverting op amp with a fixed 100k input resistor, really do have exactly 100k input impedance to within component tolerances. Many others, such as those using the also very common pattern of a 100k input resistor and 100k pot going into an inverting op amp, are close; that particular circuit varies from 100k to 125k depending on the position of the pot. Some exceptions exist, including some that are drastic (like 1k or 10M instead of 100k), which is why I said "typical" and not "always." But saying 100k is "typical" remains true.

Eurorack is designed around the assumption that input impedances are "high" and output impedances are "low" and as long as there's a big gap between those two it doesn't matter exactly what the values are. The point is that we can describe signals as voltages without thinking hard about current and power. That remains true if the difference is between 1k and 100k, or between 2k and 125k. If the gap between a typical input and a typical output is large, then the simplifying assumption that signals are pure voltages remains true even when you do things like put inputs in parallel with a passive multiple: 50k for two inputs or even 33k or 25k for three or four inputs is still enough bigger than 1k that you can get away with it. Preventing the user from needing to think too hard about impedances while patching, is a worthwhile goal. In the rare cases where really exact voltages are needed and the user can't just turn a knob to correct for the difference, then it may be necessary to think harder, but that's not so common in Eurorack, and especially, it was not the case in the specific context where I said that 100k is typical.

Is 100k a good value for input impedance? Maybe not. Would it be better to standardize on a different value, and to demand that it be more consistent among different modules? Maybe. Someone designing a new format would have the opportunity to make different design choices. I'd be interested to know what KSS is doing in his new format. And the 100k assumption is self-fulfilling and the fact it's self-fulfilling may be unfortunate. But don't shoot the messenger. Eurorack input impedances vary a lot, and are typically 100k; both those statements are true, without any contradiction.

My comment happened to be true across a huge range of situations in Eurorack, but I also made it in a very specific context, discussing why one can't freely connect a passive voltage generator with a varying output impedance of up to 60k (violating the 1k assumption) to an unspecified Eurorack input, without being likely to experience significant voltage drop. It's reasonable to calculate as I did how much voltage drop one would expect to see in such a case, using the typical Eurorack input impedance of 100k, even though some Eurorack inputs actually have other impedances. It should be understood (especially when the voltage in question is coming from a user-controlled knob) that the value won't be exact if the input impedance happens to not really be exactly 100k. The only thing "dangerous" about what I wrote was the removal of its context.

User avatar
MikeDB
Wiggling with Experience
Posts: 326
Joined: Fri Feb 08, 2019 8:28 am
Location: UK

Re: 100k input resistance... isn't

Post by MikeDB » Sun Nov 22, 2020 12:38 pm

mskala wrote:
Sun Nov 22, 2020 11:22 am
neil.johnson wrote:
Sat Nov 21, 2020 3:11 pm
I keep seeing so many posts where folks say such things as:
A typical Eurorack input has an impedance of 100k.
The more I see this, the more I realise this is such a dangerous assertion. Very rarely is this assertion actually correct, and so any discussions based on it are likely wrong from the outset.
This looks like an indirect way of calling me out for this thread: viewtopic.php?f=17&t=240512 in which I wrote those exact words. If so, it's pretty rude, and I'm not sure I should dignify it with a response. But it's true that other people have said similar things many times too.
I really don't think Neil meant it as being aimed at and certainly not rude to anybody. I think it was more the Barton attenuvertor vs the cut-down version post that had both him and me cringing.
Personally I hate cats ! Charlie the jackdaw on my shoulder also hates cats as his mother was killed by one so I had to raise him until he set off to procreate his lineage.

lickspittle
Learning to Wiggle
Posts: 21
Joined: Fri Feb 07, 2020 5:40 pm

Re: 100k input resistance... isn't

Post by lickspittle » Sun Nov 22, 2020 1:32 pm

Can I ask why anyone would not ensure a consistent input impedance? The difference between doing it right (in terms of cost, current and real estate) is vanishingly small. It would be bad enough if a mass market manufacturer cut corners to save a few pennies per unit. From a Music Tech DIY perspective, it's laughable.

User avatar
astrosound
Common Wiggler
Posts: 250
Joined: Wed Aug 06, 2014 11:14 am
Location: New England

Re: 100k input resistance... isn't

Post by astrosound » Sun Nov 22, 2020 2:51 pm

lickspittle wrote:
Sun Nov 22, 2020 1:32 pm
Can I ask why anyone would not ensure a consistent input impedance? The difference between doing it right (in terms of cost, current and real estate) is vanishingly small. It would be bad enough if a mass market manufacturer cut corners to save a few pennies per unit. From a Music Tech DIY perspective, it's laughable.
I think the short answer, as has already been mentioned above, is that so often it just doesn't matter enough to even consider. To use the example above with 100k pots hung off 100k input resistors, if I were to choose between leaving them as is or squeezing another few opamps on the board, I probably wouldn't bother. More time with layout and populating the PCB, more parts to buy, more current draw, and to what end? Besides the obvious concern with pitch CV, inconsistent input Z isn't a problem. At all.

User avatar
BugBrand
Knowledge of Bugs
Posts: 7481
Joined: Tue Jul 15, 2008 11:59 am

Re: 100k input resistance... isn't

Post by BugBrand » Sun Nov 22, 2020 3:26 pm

CLee wrote:
Sun Nov 22, 2020 9:49 am
Serge designed his modules with the protection/stability output resistors inside the feedback loop, so the varying loads are compensated for. Using banana jacks, he knew patches would stack and fan out outputs to multiple inputs. As the load increases on the output the op amp increases its gain, so as long as it doesn’t clip it stays in tune. The actual input impedance value doesn’t matter.

Too bad Euro didn’t do the same
This, absolutely - we can't still be talking about 1k output impedances, can we?
Just like people using the power lines as references - surely this got resigned to the dustbin?

lickspittle
Learning to Wiggle
Posts: 21
Joined: Fri Feb 07, 2020 5:40 pm

Re: 100k input resistance... isn't

Post by lickspittle » Sun Nov 22, 2020 4:06 pm

astrosound wrote:
Sun Nov 22, 2020 2:51 pm
lickspittle wrote:
Sun Nov 22, 2020 1:32 pm
Can I ask why anyone would not ensure a consistent input impedance? The difference between doing it right (in terms of cost, current and real estate) is vanishingly small. It would be bad enough if a mass market manufacturer cut corners to save a few pennies per unit. From a Music Tech DIY perspective, it's laughable.
I think the short answer, as has already been mentioned above, is that so often it just doesn't matter enough to even consider. To use the example above with 100k pots hung off 100k input resistors, if I were to choose between leaving them as is or squeezing another few opamps on the board, I probably wouldn't bother. More time with layout and populating the PCB, more parts to buy, more current draw, and to what end? Besides the obvious concern with pitch CV, inconsistent input Z isn't a problem. At all.
The short answer that said it didn't matter because musicians would work around the problem, had always worked around the problem so what's the problem?

socom93
Common Wiggler
Posts: 79
Joined: Sun Mar 24, 2019 1:19 pm
Location: France

Re: 100k input resistance... isn't

Post by socom93 » Sun Nov 22, 2020 4:16 pm

CLee wrote:
Sun Nov 22, 2020 9:49 am
Serge designed his modules with the protection/stability output resistors inside the feedback loop, so the varying loads are compensated for. Using banana jacks, he knew patches would stack and fan out outputs to multiple inputs. As the load increases on the output the op amp increases its gain, so as long as it doesn’t clip it stays in tune. The actual input impedance value doesn’t matter.

Too bad Euro didn’t do the same
Well a lot of people designing Euros do that too, that has nothing to do with the format, it is just linked to the designer capabilities.

Post Reply

Return to “Music Tech DIY”