Just curious if folks actually go through the hassle of measuring and pairing resistors, capacitors, and transistors used in signal paths in order to help keep everything internally balanced when assembling kits. Since many of the preconfigured kits such as Elekit tend to use 5% tolerance resistors, I was not very surprised to see large swings in resistance between the individual components. There are obviously some locations where a slight variance wouldn't be critical, like for the power supply. However, it seems that anything directly in the signal path could exhibit a balance issue if the passives had enough accumulated variance.
With OCD in full swing, when building my TU-8200, TU-8500, and most recently my TU-8550, I did take the time to value match as many pairs as possible. I don't know how much it really helped, but to my ears I don't notice any balance issues.
Anyway, just wondering if anyone else did this.
With OCD in full swing, when building my TU-8200, TU-8500, and most recently my TU-8550, I did take the time to value match as many pairs as possible. I don't know how much it really helped, but to my ears I don't notice any balance issues.
Anyway, just wondering if anyone else did this.
Should always check parts for value before soldering, it saves a lot of mistakes.
Might as well match them at the same time, it can't hurt.
Might as well match them at the same time, it can't hurt.
Depending on the function a part has in the circuit, matching may have huge impact on performance. For example, common mode noise rejection of a balanced-to-single-ended op-amp based convertor depends to a large degree on resistor matching.
Yes, and you want actual precision parts for a circuit like that, as looser tolerance parts
also have looser specs on drift, etc.
also have looser specs on drift, etc.
In general, save for measuring equipment, competent designs work fine with 5% parts tolerance.
Where precision is needed, as in biasing, usually an adjustment trimmer is available.
And when a multicomponent gain block is involved, such as in Preamps and Power Amps, you only really need to match TWO NFB resistors per stage.
And even if "horrible" 5% tolerance resistors were used there, allowing up to 10% combined error, level difference would amount to 0.9dB
Where precision is needed, as in biasing, usually an adjustment trimmer is available.
And when a multicomponent gain block is involved, such as in Preamps and Power Amps, you only really need to match TWO NFB resistors per stage.
And even if "horrible" 5% tolerance resistors were used there, allowing up to 10% combined error, level difference would amount to 0.9dB
If every stage of a system has just 1dB imbalance, that could add up to a problem level overall.
They might cancel somewhat, or they might not. And 1% resistors are not a major cost in a system.
They might cancel somewhat, or they might not. And 1% resistors are not a major cost in a system.
Last edited:
Yep, I do that occasionally. Here is one example (link to thread). The final application circuit is presented in post #32 of that thread.
I have rebuilt many Carver C-9 Sonic Hologram units over the years. I wanted to get the most of these units as I could by using the highest quality parts that could be afforded. In a couple of cases, I replaced every single possible part. Only the circuit board was original. Cutting to the point, these two units sounded not better or worse than other units. Even more interesting to me was that yes, some units sounded slightly better than others. It seemed that less than my input of parts, something not known to me made certain units simply sound more live than others. I have no issues with using the best parts, yet I do not have definite proof of noticeable sonic improvement of 'upgraded units'. FWIW it still doesn't stop me from making each to the best of my ability, and maybe the design itself lends itself to a less than perfect parts list.