Sorry I thought I was responding to the other thread haha, and thank you
.
600 Ohm inputs
- Thread starter Brian Roth
- Start date
Help Support GroupDIY Audio Forum:
.user 37518
Well-known member
I am definitely not crazy about the line inputs just being padded and then fed to the same discrete stage of the pre-amp, neither for the design of that pre. If this is being used solely for line-level signals in a studio (i.e., as an audio interface), I would definitely bypass the whole input stage. I am trying to find where X39 and X40 go to, I am guessing they go to some op-amp inputs just like the ones shown next to these outputs. I don't see any global feedback in that input stage, just the Sziklai diff pair going out, nothing returning; not the best design if you ask me, looks like a low-budget prosumer preamp from the 80s.
Before even thinking about adding some input/output transformers, I would instead fix that input stage first.
Some time ago I bought a Behringer UMC404HD interface to perform some measurements, so I don't know if it sounds ok or not, neither the distortion levels, but, if the input stage is the same as this, I am afraid I made a bad purchase.
Last edited:
Would an OPA1632 be any good as a buffer for driving a 600 ohm device?
My audio interface is struggling to drive my Gyraf Pultec build so I'm wondering about a buffer box as a fix. Would something like this (from sound-au.com) work? There's a similar circuit on the datasheet.
My audio interface is struggling to drive my Gyraf Pultec build so I'm wondering about a buffer box as a fix. Would something like this (from sound-au.com) work? There's a similar circuit on the datasheet.
To what level? The datasheet has distortion numbers for a 600 Ohm load, but then a lot of the specs are noted as 800 Ohm load, and even for 800 Ohm load the maximum output swing is reduced from a 2k or greater load.
The THD vs output voltage graph has a line for 600 Ohm load, and it looks really reasonable up to around 22 dBu, but it is rising pretty quickly at that point. It looks like it can't quite get to 24dBu output with 600 Ohm load.
You would need to pay a lot of attention to the power and thermal limits at that load as well.
The Gyraf schematics I have seen suggested a Lundahl input transformer that worked OK with a 10k terminating load. It seems like it might be easier to just swap the input transformer and resistor to have a higher input impedance than trying to build a buffer box to stick in front of your Pultec clone (assuming you even need to swap the transformer, you would have to check the spec sheets to see how it performs with a higher secondary termination).
Matt Syson
Well-known member
As soon as you divert from the original concept of 600 Ohm interconnections things will fall apart in a variety of ways. 600 Ohm source impedance into 600 ohm 'load' is significant and you have to take on the differences in 'originality' to simple functionality. Of course EVERYTHING done in the pasthas disappeared almost completely. equipment has aged, peoples ears have aged, everything is wrong and cannot be reproduced. Even down to having a tiff with your 'significant other' before starting a mixing session will affect you. I made a mistake with a Pultec copy and used a higher impedance transformer. 1:1 ratio but intended for higher impedance rather than 600 Ohm line level. The cut and boost was all wrong as the whole;e circuit (passive) depends on known source and load impedances.
Padding an IC based output circuit to give a 600 Ohm 'send' impedance is still not the same as sending from a transformer in the previous piece of gear, and in turn how this is driven. So just the input to a Pultec depends on 'authenticity' of the source (sending) output and it's previous stage, never mind what the output of the Pultec is actually feeding. which should be 600 Ohms and posssibly with a reactive element (transformer not just resistive. Chasing 'authenticity' with a complex system that was a product of the specific time and environment is a crazy idea. Like saying the combatants in the American Civil war should have fought harder.
Padding an IC based output circuit to give a 600 Ohm 'send' impedance is still not the same as sending from a transformer in the previous piece of gear, and in turn how this is driven. So just the input to a Pultec depends on 'authenticity' of the source (sending) output and it's previous stage, never mind what the output of the Pultec is actually feeding. which should be 600 Ohms and posssibly with a reactive element (transformer not just resistive. Chasing 'authenticity' with a complex system that was a product of the specific time and environment is a crazy idea. Like saying the combatants in the American Civil war should have fought harder.
Yes my understanding is that any change to the transformer will affect the filter circuitry because of the way it interacts. This is something Jakob has discussed in the past so I'm hesitant to make a change that would severely affect the circuit behaviour.
According to the specs, the output impedance of my interface is 300 ohm and it can output up to 19 dBu - a bit lower than ideal, but comfortably within the spec of the chip I think?
What effect can I expect it to have on the signal? When driven from my interface, there is an increasing low end rolloff from 10k until it's 3dB down at around 100hz, and the rolloff continues to increase below that. If I drive it from my TG1 with a 600 ohm output transformer, performance is vastly improved - gentle rolloff starts below 50hz. However, I don't have a 600 ohm device to drive from the EQ so that goes straight back into my interface.
I appreciate it's not an ideal setup, but I'm hardly the first person to attempt to use 600 ohm gear in a modern studio. I admit I was naïve in thinking it would 'just work' and I've since learned a lot about impedances. I don't mind if it's not 100% perfect - I was just hoping for a simple solution that wouldn't result in having to spend thousands on a new convertor.
Matt Syson
Well-known member
So much of 'modern' gear is NOT designed and tested to provide a true 600 Ohm output impedance with corresponding voltage headroom to overcome the 6dB loss in level when you actually place a 600 Ohm load on the output. Much of it presumes that the load will be in the order of 10K Ohms or more (previously referred to as a bridging load) where gear in the days of valves with a 600 Ohm output impedance could drive up to 10 or 15 '10K bridging loads before the drive capability (distortion etc) was compromised. People born after the age of valves have not bothered to learn WHY so many audio interfacing standards came about, and the sometimes serious implications. Still, Hi Fi was about 50 to 10KHz bandwidth at 'less than about a couple of percent distortion' Variously 10% then more stringent 1 or 0.1 % as technology improved. The LA2A compressors start to run into significant distortion (by today's terms) when delivering more than +18dBu into a 600 Ohm load unless the output valve is in good condition. Running into 10K (typical convertor or mixing desk input) it is happy to above +24dBu if I remember correctly.
As a handy aside ALL the revered 'Neve' gear is NOT correct. The types of capacitors and other components of the originals have gone (decomposed) so NO modern version can get back to the space and time that existed when the 'classic' recordings were made and then you have to considered that aLL old recording mediums are 'dead' (have moved on and been fitted with new parts simply to make them work at all). True 600 Ohm outputs and inputs are a nightmare to wotrk with because you simply can't just put another 600 Ohm load in parallel because you lose another 3dB (IIRC) so in places where this was a likely requirement you had Distribution amplifiers with designed in capability to actually feed typically up to 10 X 600 Ohm loads. Broadcasters had racks full of distribution amplifiers so essentially unity gain but ability to drive sufficient current into the extra loads NOT wired directly in parallel because each of the 10 outputs has it's own resistors/transformer winding to keep them all separate. the premise being that you could accidentally short any one or more of the 10 outputs and the signal would not be impaired. Even having a 'listen' jack on a patchfield where someone could plug in a pair of what were often 600 Ohm headphones meant a possible compromise that is difficult to circumvent.
As a handy aside ALL the revered 'Neve' gear is NOT correct. The types of capacitors and other components of the originals have gone (decomposed) so NO modern version can get back to the space and time that existed when the 'classic' recordings were made and then you have to considered that aLL old recording mediums are 'dead' (have moved on and been fitted with new parts simply to make them work at all). True 600 Ohm outputs and inputs are a nightmare to wotrk with because you simply can't just put another 600 Ohm load in parallel because you lose another 3dB (IIRC) so in places where this was a likely requirement you had Distribution amplifiers with designed in capability to actually feed typically up to 10 X 600 Ohm loads. Broadcasters had racks full of distribution amplifiers so essentially unity gain but ability to drive sufficient current into the extra loads NOT wired directly in parallel because each of the 10 outputs has it's own resistors/transformer winding to keep them all separate. the premise being that you could accidentally short any one or more of the 10 outputs and the signal would not be impaired. Even having a 'listen' jack on a patchfield where someone could plug in a pair of what were often 600 Ohm headphones meant a possible compromise that is difficult to circumvent.
When you design your next console you can add as many active line inputs as you want to every strip.
It is fairly common in modern console design to send the different inputs into the one active gain stage (cost and PS current consumption).
While a simple line input might be slightly higher performance, the net effect on the audio quality of existing line input sources is generally not significant.
JR
user 37518
Well-known member
This is not a console, John, this is an 8 channel audio interface with a crappy front end. That is basically all the analog portion of it. Bad cheap design.When you design your next console you can add as many active line inputs as you want to every strip.
It is fairly common in modern console design to send the different inputs into the one active gain stage (cost and PS current consumption).
While a simple line input might be slightly higher performance, the net effect on the audio quality of existing line input sources is generally not significant.
JR
SM Pro Audio have a ready made 8 channel box - the EP-84 - which has 47KΩ input impedance with mic and line balanced inputs going to a balanced 600Ω output. Saves inventing wheels. Cheap too.
On the more expensive side of things D.A.C.T. have the CT101 modules which although unbalanced could drive an output transformer to balance the output
These can be ganged up to the number of outputs required
On the more expensive side of things D.A.C.T. have the CT101 modules which although unbalanced could drive an output transformer to balance the output
|
Gavin Pursinger
Well-known member
Gavin Pursinger
Well-known member
Thx everybody for this discussion...I love the 600 Ohm world and all the glowing bottles...this has explained alot of what I hear from my builds with 500 or 600 Ohm outs to modern inputs and console send,returns. I guess I'll hhave to live with it as I don't have enough knowledge to remedy it or build a time machine ...thx to.all you very knowledgeable cats
Which of course is unrealistic.
I suggest the persons who operate the equipment are educated about the limitations pertaining to some combinations of gear and recommand the use of a few dedicated adapters, either accessible from the patchbay or permanently attached to questionable equipment.
Basic specs would be:
- unity gain
- input Z >10k
- output Z <50 ohms with possibility to adjust up to 600
- max in/out level +24dBu
- adequate FR and distortion performance
Matt Syson
Well-known member
The max output level to achieve +24 dBu into 600 Ohms but itself presenting an output impedance of 600 Ohms is not realistic with a 1646 because the output stage would have to provide +30dBu (for resistive output impedance build out) or some cunning way to make the near zero output impedance of these chips 'appear' to be 600 Ohms. The TubeTech range of gear outputs are specified to give +24 /28 dBu into 600 Ohms and with new output valves will give around +34 to +36dBu into 10K ohms or greater. So a 1646 will give about +26 dBu into a 600 Ohm LOAD BUT to add a passive resistor network to make the send impedance LOOK like 600 Ohms source (2 X 270/300 Ohm resistors) will drop the maximum output voltage at clipping to about +20dBu. It is the passive EQ units (Pultec type design) that NEED a true 600 Ohm SOURCE impedance to make them perform to the original design specifications. Comp/lim units would not be so fussy as their input impedance is basically not frequency and impedance dependent.
My suggestion was based on operators being adequately educated, so they would understand the nedd to perhaps use one such unit at the input and output of concerned device.
True. However, I don't think they need to operate with actual +24dBu headroom to operate correctly.
Under most circumstances, my suggestion would be a workable practical solution.
Actually, I sold thousands of such devices that had only +22dBu open circuit headroom, with < 20 ohms output Z performing such tasks. Nobody ever complained.
Even including 6dB loss due to matching.
Following from my previous post (bass rolloff when connecting my audio interface to my G-Pultec), I decided to breadboard up a version of this circuit, using 10k resistors instead of the suggested 5k6, and 600 ohm output resistors. I used a 5532 running off +/-12V, which I figured would be enough for test purposes. I omitted C101 and C102.
It still suffers from the low end roll off issue. Initially I thought it was because I should have used 300 ohm output resisters - but this brought the roll off point up to ~5kHz.
Please could someone explain why this doesn't work? Can it be modified to work?
It still suffers from the low end roll off issue. Initially I thought it was because I should have used 300 ohm output resisters - but this brought the roll off point up to ~5kHz.
Please could someone explain why this doesn't work? Can it be modified to work?
What happens if you replace r109 & R110 with 10 ohms?
Your G-Pultec has an input xfmr, right? What is it?
Yes it does, a Lundahl LL5402 (600 ohm).
I’ll try the lower resistors - my results do suggest that going the other way in value would move the cutoff down.
Datasheet says: ideally used with mixed feedback drive circuits
It means it has low nominal inductance and should be driven from a very low source impedance. Even with zero source Z, it will never be optimized. It really needs the active negative source Z to deliver the goods.
That sounds like it was designed as an output transformer, not for input.
Yes it is an output transformer.
Last edited:
Cranehazard
Well-known member
- Joined
- Feb 3, 2016
- Messages
- 54
My approach is, I build all my own preamps tube and solid state. I make sure they can put 25v to 30v point to point into a 600 ohm load. I have various interfaces. I was surprised to find the supposed 10k inputs acted like 600 ohms. It takes 20v point to point to clip my motu m4 line inputs. To truly see output impedance you need to look at the output with and without a load if the signal sags more than like 5% when you attach the load the output impedance is too high. Lower output impedance is always better. If your equipment can't drive a 600 ohm load without the signal sagging its not good gear. Almost any audio op amp can drive 600 ohms and tube gear can be designed to drive 600 ohms and lower. If all your gear has good output drive capability the whole system will sound better. Just my method.
