Worrying about travel to a place before some aspect of the area changes is a loosing battle. I went to Stonehenge before it was fenced. I climbed on the stones... So now that can't be done - What ya' gonna' do? It's like saying "I wish I went to Hawaii before it was a state." It's changed, that's it! - No more wild weed growing on the foothills free for the taking. Life goes on.
I was feeling quite uneasy that I was the only one to report transistors singing.
But now three participants (jneutron, PMA, RNMarsh) confirm it.
I’ll stop looking my face on the mirror watching for signs.
I guess their experience was with the devices on the bench under high drive.
Pavel , ref “necessary condition”
Fixing the device to a metal heatsink, will have some effects:
1. Lowers the frequency of vibration due to added mass, i.e. shift noise spectrum toward frequencies that ear is more sensitive to.
2. Increases the vibrating surface, making sound emission more efficient.
3. The edges of the heatsink cooling fins emit mid to high frequency sound very well.
Some damping from the aluminum (material of the heatsink) itself can be provided.
I find that instead of using silicon paste, the elastic thermal pads between the device and the heatsink, in combination with attaching bolts made of soft material (bronze or plastic instead of steel) and not much torque, provide needed additional damping.
Care on the area where the heatsink attaches to the metal case of the equipment is important. I apply a bead of silicon thermal conductive pad at the mating surfaces before fastening.
In addition to these, rounding off all the fin edges with a fine file and applying a bead of hot silicon glue there, also helps.
Normal? What is normal?
Do you remember the scene on the movie with Igor and Dr. Frankensteen?
George
But now three participants (jneutron, PMA, RNMarsh) confirm it.
I’ll stop looking my face on the mirror watching for signs.
I guess their experience was with the devices on the bench under high drive.
Pavel , ref “necessary condition”
Fixing the device to a metal heatsink, will have some effects:
1. Lowers the frequency of vibration due to added mass, i.e. shift noise spectrum toward frequencies that ear is more sensitive to.
2. Increases the vibrating surface, making sound emission more efficient.
3. The edges of the heatsink cooling fins emit mid to high frequency sound very well.
Some damping from the aluminum (material of the heatsink) itself can be provided.
I find that instead of using silicon paste, the elastic thermal pads between the device and the heatsink, in combination with attaching bolts made of soft material (bronze or plastic instead of steel) and not much torque, provide needed additional damping.
Care on the area where the heatsink attaches to the metal case of the equipment is important. I apply a bead of silicon thermal conductive pad at the mating surfaces before fastening.
In addition to these, rounding off all the fin edges with a fine file and applying a bead of hot silicon glue there, also helps.
Normal?
What is normal?Do you remember the scene on the movie with Igor and Dr. Frankensteen?
George
Last edited:
jcx
Thermal expansion (not reciprocal) is another factor of signal - related vibration transmission.
Piezoresistivity (not reciprocal) is another microphony factor to consider.
Thank you for helping feeding the pile.
Crystalline silicon (centro-symmetric) is not piezoelectric. I could not find if doped silicon (P, N lands) is piezoelectric. I asked Scott and in plain words he said yes.
Plain words I understand.
I can’t understand though if Scott means piezoelectricity too here:
George
So, both you (acceleration) and Scott (force) agree (F=ma) that this tinny mass can’t do it or be affected by it to any reasonable extent. I am confident you know enough to say so.
Thank you for the explanation. This can apply also to wirewound emiter resistors of output transistors, no?
I risk to say that Lorenz force fits as an answer to question no. 1 of the post linked below.
Can you have a look please and comment when time permits?
And now it’s a good time for some cascaded speculative questions.
George
In most cases it's strain guage effects, i.e. the channel of a FET has a change in Idss or a diffused resistor gets a delta R. There is a wonderful paper from RCA with a 3D mathematical (not simulation) solution for one of their first NTSC color matrixing chips. I have a copy buried somewhere, you would really like reading it if this topic is of interest. The stress between die and lead frame of some die mounting processes (after cooling) are quite substantial.
jneutron is right, imparting actual strain on a die by shaking the package would require enormous G forces. I hope I have correctly communicated the difference between stress and strain, it is very important. The strain (the actual bending of the die in response to stress) is what is important.
Last edited:
Yes some of the 1000 yr. old masterpieces at the MFA in Boston were at one time folded. You have to realize that the original work is on "rice" paper (actually a very special mulberry bark) it is very thin and you (believe it or not) completely wet the entire painting and relax it before mounting it on backing paper with pure wheat starch paste. Not a single artifact of the folding is visible after drying, though at the midpoint of the process the entire painting is incredibly vulnerable to total destruction. The guys in the back room using flat irons and dry mounting paper you see now is just "modernization".
EDIT - Shanghai and Beijing are Analog offices, never look a gift horse in the mouth, just fill it with stinky tofu.
I designed in that part together with the AD7672 12bit ADC back in the mid 80s. Had a little curiosity that it came in a ceramic package - was that intentionally chosen for lower DA of the parasitic?
Word on the street is that the best stinky tofu is to be found in Shaoxing, about an hour's bus ride from here.
Last edited:
You guessed it, I hear the stinky tufu tripe sandwitch is "real man's" food. IIRC Zimmern could not get it down, and he was in your neighborhood. Didn't phase me, nor does natto.
The key word here is "directly", it's the indirect ways that cause the audible problems. Meaning, that if the disk is made in, physically subtle, different ways from another, digitally identical, disk then the CD player will have to work harder electrically, or differently, to cope with one disk versus the other. This will mean that the pattern of current draw pulses from the power supply will vary, altering the the level and spectrum of interference effects upon the analogue circuitry within the player and likely also impacting connected equipment.
Unless, of course, you believe in perfect power supplies and the distribution of their product ...
Frank
In my opinion you can't say for sure whether its poor engineering unless you know what the original constraints were. 'Was the customer still happy with the result, at the price?' is not mentioned here for example. IMO over-engineered is just as much poor engineering as under-engineered.
Last edited:
I would suggest that there's engineering, and then there's engineering. In the first instance, that which is sufficient to meet accepted standards of performance, and then a "higher" level which can produce a performance, which for those people who are sensitive to those subtleties, is superior. The sort of thing that "audiophiles" have been ranting on about for ages ...
The analogy with car performance is obvious, a good engineer will design a vehicle that is competent at 100mph, that will satisfy most people. Yet, some drivers will yearn for one that can be flung around in outlandish ways at those sorts of speeds, and be proficient doing so. A notch up in attention to detail is necessary for that ...
So, IMO, it is not a case of good vs. poor engineering; rather that which is "good enough" in the commercial world, and that which is aiming to deliver the goods no matter what sort of test is applied to it ...
Frank
Last edited:
Well abraxalito, I can understand that you can attempt to fault the CTC Blowtorch as to being 'over-engineered' and perhaps we could have saved some money ignoring one thing or another, hoping that it would not be missed. However, in MY experience, I don't know what to compromise AND get the same sonic results.
We could have made it with a cheap steel case that looked almost the same on the surface, used IC's (good ones) and relays. Many do that, already, yet do they get the same results? Maybe close enough for you! '-)
We could have made it with a cheap steel case that looked almost the same on the surface, used IC's (good ones) and relays. Many do that, already, yet do they get the same results? Maybe close enough for you! '-)
- Status
- Not open for further replies.