All well and good! Nice video, and nice topic! Just one thing to have in mind. Output impedance is just that, an impedance and not a resistive / fixed resistance, thus it can and usually do vary with frequency. Looking at both the phono pre and both the universal and e80cc pre, they have a 2.2uF output coupling cap which is quite typical of an output stage. When the output impedance was measured , it seems that was at 1Khz. At this frequency the impedance of the output cap is 72 ohm, so just around 10% of the total output impedance. If we move down in frequency we see at 100Hz the impedance of the cap is now 723 ohms, so it will have almost doubled the output impedance of the stage at this frequency, so we would be now in the area of 1.5 K ohm output impedance. Moving down to the low B of a 5 sting bass guitar at 31 Hz we get 2333 ohms of the cap alone, bringing the total output impedance to around 3k. And at the nominal lowest frequency 20Hz used for stating the frequency response , we have 3617 ohm in the cap alone bringing the total output impedance at that frequency up to over 4k. So loading that output impedance with a 1K input impedance will leave 3/4 of the bass signal in the output impedance. I believe you should see a significant bass roll-off when you do a frequency response plot of one of the preamps into a 1k load?
Interesting, and yes as per industry standard we test at 1Khz. We didn't notice a bass roll off, probably because most vinyl, even good vinyl records are rolling off higher anyway. Currently we have the stepped attenuator located after our Universal 6or12Sn7 Preamp, so feeding into monoblocks with an input impedance of 470k. In the future we'll include a 40hz, 1khz and 10khz output impedance value, based on a proper bench test of the actual circuit. It will be interesting to see if we can confirm your data. Thanks!
@@tubelab194 Will be interesting to see if you see the same yes. I made myself a little test jig today so I could easily switch between 47k, 18k, 10k and 1k loads. Then used these load on the phono pre, and used the Analoge Productions test LP that has a 1k down to 20hz sweep as one of the tracks. Measured with scope and mv meter. Found no issue down to 10k load but at 1k the signal all over was halved as expected since we basically have same load as output impedance and thus have a voltage divider. That may be fine if you have enough gain in the system anyway. However here we were also good frequency response wise down to 100hz but below this it started to drop of. Dropped about 5.8 db from 100Hz to 20Hz.
You are correct, bench testing confirms the low frequencies start to roll off at about 100hz, it looks like a first order slope, so not very fast, but still not a good thing - in our design work, we fight hard for every low frequency db we can get! We've designed a new Set "B" with an input resistor of 10k, hopefully as a better standard input impedance. Let me know what your modelling shows as the upper limit for an input device with a typical 2.2uf output capacitor and I'll see if we can confirm. BTW thanks much for your observation, it will definitely make an improvement to what is already a great sounding volume control and in future we'll include bench testing specifications that include 20hz 1Khz and 20Khz, so this kind of design error doesn't slip past us again.
@@tubelab194 Great, well as you mentioned on the other channel, that the modeling may not be too accurate for simulating tube output impedance seems to be right. The simulation showed clear roll-off at 10k load, but my measurements / tests did not show this at 10k. So it seems the limit is somewhere between there and needs to be determined by testing.
@@RaviNewfarm yes bench testing is always the last arbiter, well listening tests as well, but listening can't always detect a first order roll off on the edge of the frequency band. I'm building a new set based on a 10k input resistor. It should give us plenty of room for any equipment with an output impedance of 1k or less. This is actually the Set "B" that I had planned on making, but I think it will end up being the standard everyone uses.
No, it would put everyone to sleep, especially when we do the math portion. I don't even enjoy it, and I like testing stuff, it's just something we have to do to complete our Kit Amp specifications - we already know we're low impedance by design, jist not the exact value.
I just cant easily get my head around resistance. It feels to me that if a low R is coming OUT (let's say 500 Ohms) into a higher resistance IN (75k ohms), energy is being STOPPED because of the higher resistance IN…But science is saying the opposite. IS there a better way of explaining to someone like me? Thank you
That's exactly correct and I agree, until you get deeper into understanding how a circuit actually works at the electron level, it's hard sometimes. And in my early days of design some things I just had to accept. Think of it this way, if the signal was heading down a wire with infinite room for it go, it would be dissipated - right? So by creating a nice bit of resistance, say 1:10 minimum, that signal isn't being dispersed or dissipated. It's much more complicated than that because we have other passive electric components (capacitors and resistors) that are often in series and or parallel. And those other components can create filter networks and resistance. Even linking to the previous stage! So designers came up with a standard rule of a 1:10 minimum impedance ratio and preferably 1:100, with a 1:10 ratio you basically avoid almost all of the potential negative interactions. Note, measuring input impedance is easy, jist follow the circuit and the lowest resistance to ground is the input impedance. Output impedance is much more involved and requires an electrical bench and a specific set of procedures to calculate it. Luckily every properly designed bit of gear will clearly specify the input and output impedance. And we of course provide that data on our spec sheets for all of our Kits.
I going to have to agree to disagree (based on extensive testing and listening in a controlled 2 channel treated room. Also many years of electronics knowledge). My rule of thumb is Out/in should be at minimum 100X and 1000X for high resolving systems
@@audiononsense1611 in most circuits 10x will work just fine with no signal loss. We are both audiophiles and work professionally as designers, and as a result we spend a lot of time listening to our prototypes as well as bench testing the gear. The real problem has been the number of poorly thought out and poorly designed bits of gear, particularly out of China, though we've seen some terrible designs from everywhere including Europe.
All well and good! Nice video, and nice topic! Just one thing to have in mind. Output impedance is just that, an impedance and not a resistive / fixed resistance, thus it can and usually do vary with frequency. Looking at both the phono pre and both the universal and e80cc pre, they have a 2.2uF output coupling cap which is quite typical of an output stage. When the output impedance was measured , it seems that was at 1Khz. At this frequency the impedance of the output cap is 72 ohm, so just around 10% of the total output impedance. If we move down in frequency we see at 100Hz the impedance of the cap is now 723 ohms, so it will have almost doubled the output impedance of the stage at this frequency, so we would be now in the area of 1.5 K ohm output impedance. Moving down to the low B of a 5 sting bass guitar at 31 Hz we get 2333 ohms of the cap alone, bringing the total output impedance to around 3k. And at the nominal lowest frequency 20Hz used for stating the frequency response , we have 3617 ohm in the cap alone bringing the total output impedance at that frequency up to over 4k. So loading that output impedance with a 1K input impedance will leave 3/4 of the bass signal in the output impedance. I believe you should see a significant bass roll-off when you do a frequency response plot of one of the preamps into a 1k load?
Interesting, and yes as per industry standard we test at 1Khz. We didn't notice a bass roll off, probably because most vinyl, even good vinyl records are rolling off higher anyway. Currently we have the stepped attenuator located after our Universal 6or12Sn7 Preamp, so feeding into monoblocks with an input impedance of 470k. In the future we'll include a 40hz, 1khz and 10khz output impedance value, based on a proper bench test of the actual circuit. It will be interesting to see if we can confirm your data. Thanks!
@@tubelab194 Will be interesting to see if you see the same yes. I made myself a little test jig today so I could easily switch between 47k, 18k, 10k and 1k loads. Then used these load on the phono pre, and used the Analoge Productions test LP that has a 1k down to 20hz sweep as one of the tracks. Measured with scope and mv meter. Found no issue down to 10k load but at 1k the signal all over was halved as expected since we basically have same load as output impedance and thus have a voltage divider. That may be fine if you have enough gain in the system anyway. However here we were also good frequency response wise down to 100hz but below this it started to drop of. Dropped about 5.8 db from 100Hz to 20Hz.
You are correct, bench testing confirms the low frequencies start to roll off at about 100hz, it looks like a first order slope, so not very fast, but still not a good thing - in our design work, we fight hard for every low frequency db we can get! We've designed a new Set "B" with an input resistor of 10k, hopefully as a better standard input impedance. Let me know what your modelling shows as the upper limit for an input device with a typical 2.2uf output capacitor and I'll see if we can confirm.
BTW thanks much for your observation, it will definitely make an improvement to what is already a great sounding volume control and in future we'll include bench testing specifications that include 20hz 1Khz and 20Khz, so this kind of design error doesn't slip past us again.
@@tubelab194 Great, well as you mentioned on the other channel, that the modeling may not be too accurate for simulating tube output impedance seems to be right. The simulation showed clear roll-off at 10k load, but my measurements / tests did not show this at 10k. So it seems the limit is somewhere between there and needs to be determined by testing.
@@RaviNewfarm yes bench testing is always the last arbiter, well listening tests as well, but listening can't always detect a first order roll off on the edge of the frequency band. I'm building a new set based on a 10k input resistor. It should give us plenty of room for any equipment with an output impedance of 1k or less. This is actually the Set "B" that I had planned on making, but I think it will end up being the standard everyone uses.
do the output impedance demonstration please. would really like to see it.
No, it would put everyone to sleep, especially when we do the math portion. I don't even enjoy it, and I like testing stuff, it's just something we have to do to complete our Kit Amp specifications - we already know we're low impedance by design, jist not the exact value.
@@tubelab194 some of us like the math!
It looks a lot like a tax stamp on the bottom of cigarette packs. At least here in Minnesota.
I just cant easily get my head around resistance. It feels to me that if a low R is coming OUT (let's say 500 Ohms) into a higher resistance IN (75k ohms), energy is being STOPPED because of the higher resistance IN…But science is saying the opposite. IS there a better way of explaining to someone like me? Thank you
That's exactly correct and I agree, until you get deeper into understanding how a circuit actually works at the electron level, it's hard sometimes. And in my early days of design some things I just had to accept. Think of it this way, if the signal was heading down a wire with infinite room for it go, it would be dissipated - right? So by creating a nice bit of resistance, say 1:10 minimum, that signal isn't being dispersed or dissipated. It's much more complicated than that because we have other passive electric components (capacitors and resistors) that are often in series and or parallel. And those other components can create filter networks and resistance. Even linking to the previous stage! So designers came up with a standard rule of a 1:10 minimum impedance ratio and preferably 1:100, with a 1:10 ratio you basically avoid almost all of the potential negative interactions. Note, measuring input impedance is easy, jist follow the circuit and the lowest resistance to ground is the input impedance. Output impedance is much more involved and requires an electrical bench and a specific set of procedures to calculate it. Luckily every properly designed bit of gear will clearly specify the input and output impedance. And we of course provide that data on our spec sheets for all of our Kits.
@@tubelab194 I guess I'm just going to accept it. But I'd like to intuitively understand it. Thank you for the added ideas...
I going to have to agree to disagree (based on extensive testing and listening in a controlled 2 channel treated room. Also many years of electronics knowledge). My rule of thumb is Out/in should be at minimum 100X and 1000X for high resolving systems
@@audiononsense1611 in most circuits 10x will work just fine with no signal loss. We are both audiophiles and work professionally as designers, and as a result we spend a lot of time listening to our prototypes as well as bench testing the gear. The real problem has been the number of poorly thought out and poorly designed bits of gear, particularly out of China, though we've seen some terrible designs from everywhere including Europe.