Tubes vs. transistors & quantization distortion

Scifi

Super Member
Tubes vs. transistors & quantization distortion

After reading an article in a 1987 issue of Discover magazine about the popularity of tube amps I decided to try building one using TV tubes and then later, HiFi tubes (and still later, I collected some tube monoblocks and stereo amps I found in antique and thrift stores). Before this I was experimenting with solid state audio circuits and tubes seemed to give me what I was looking for. One day while pondering the tube vs. transistor sound mystery I tried visualizing the inner workings of tubes and transistors. With tubes, amplification seems to take place (almost) entirely in a vacuum and classical electromagnetic theory seems adequate to explain most of it. With transistors you basically have to start with quantum mechanics to explain how they work. An idea came to me, could it be possible that solid state electronics produce a kind of quantization or quantizing error distortion because of their quantum mechanical nature? It seemed to me solid state sound had an unnatural characteristic similar to the digital sound from CDs. I wrote to Ed Dell in 1990, Amateur Audio Publications, Inc. editor, about this idea and he published my letter, in Glass Audio magazine and he wrote back and told me about some publications by Malcolm Hawksford about some work on “the molecular and atomic activity within solid state devices as well as work on quantizing problems”. I looked through some books in the library about transistors and didn’t find much that would lead me to the answer. I did find something about traps in semiconductors and a graph that had steps in the audio region, but don’t know if that’s related or not. After finding Ed Dell’s 1990 letter recently while looking through some old papers, I decided to look on the internet to see if there are any articles by Hawksford available. I found a site with a collection of them and one article of his which discusses quantizing distortion in transistors. Looking through the equations, it appears to me that it doesn’t have anything directly related to quantum mechanics in them. The website is
http://www.essex.ac.uk/csee/research/audio_lab/malcolms_publications.html
and the article I read is
J7 FUZZY DISTORTION IN ANALOG AMPLIFIERS: A LIMIT TO INFORMATION TRANSMISSION?, M.O.J. Hawksford, JAES, vol.31, no.10, pp.745-754, October 1983
http://www.essex.ac.uk/csee/research/audio_lab/malcolmspubdocs/J7 Fuzzy distortion.pdf

I’m wondering if this effect is seen in tubes or not. If it is, maybe it’s a lot less compared to transistors based on their subjective sound qualities. Could quantizing distortion be the main cause of the difference we hear in tubes and solid state circuits? Maybe its only seen/heard in solid state and not in tubes to any significant degree if at all.
 
Last edited:
OK...no one followed up, so maybe I scared them off.

My problem is this: Any paper that starts off with a premise such as "Tubes sound better than solid state, and I shall attempt to ferret out why in this document" is flawed from the get-go. He dances around the direct statement that transistor amplification is inherently flawed (we've all heard good and bad from each), but leads the reader to the door and hands him the key to the room with his own conclusions.

I read that paper years ago, and since then the argument about TID has been shot dead as Lincoln (for two decades or more); thus, a large part of the basis for his argument looks pretty dated now.
 
probably not more than a handfull here that TRULY understand the document... and I'm not one of them. :)

wished I did though!!
 
I had someone telling me "tubes are best" or "tubes are better than transistors". I like tubes but can't deny that I like transistors too. I like the tube sound in some of my songs that sounds .... what's the word.... magical? in some passage that none of my solid state receivers were able to produce. Other hand, solid state produces the sound I wanted the tubes does not produce. It would depend. I like both worlds. Some like transistors fine, tubes fine, what is important is your own enjoyment.
 
There is no single ultimate means of amplification. Any device can and will introduce its own amount of distortion. You just have to pick the distortion that works best for you.
 
I dont descriminate between tubes and transistors...they both get love from me. It just depends on which speakers Im running.
 
Tubes also use quantized charge carriers small enough to be strongly affected by the effects of quantum mechanics, electrons. A single electron (released by the strike of a single photon) can easily be detected by a photomultiplier tube.

The first section of the article mentions many possible distortion characteristics in transistor circuits, probably more than I've seen in any article, and some I haven't heard before (such as the capacitance between base and collector changing as Vbe changes), but the fact that I've not seen many of these addressed elsewhere suggests to me they're not significant, compared to other, more well-known and clearly-estabilshed effects. There are other more practical analyses of distortion, such as Douglas Self's "Distortion in Power Amplifiers" page.

Self has several books on solid-state audio electronics design, as do other authors, including a new one by Bob Cordell as discussed over on diyaudio.com. I'm "almost" interested enough to get these, but there's only so much time in life and I already have too many projects and interests.
 
Although TID, crossover distortion, overhang and NLDM is discussed in the beginning of the paper, the paper is NOT about any of these types of distortion. The paper is about an extension of West's and Curtis's theories (references 13 and 14). Hawksford seems to use Beaufoy and Sparkes's theory (reference 15) about base stored charge and amplification and comes up with a theory about a limit to ultimate resolution which appears to have some kind of quantization distortion similar to pulse code modulation. I'm trying to read Hawksford's article again more carefully. It looks like reading references 13, 14 and 15 would be a big help in understanding it, so I'll have to get copies of those sometime.
 
Last edited:
You DO need quantum mechanics to account for all aspects of vacuum tube
operation, especially the phenomenon of thermoionic electron emission from
the cathode. Welcome to AK Scifi, you picked your name well...
 
I got copies of references 13, 14 and 15 and scanned 13 and 14. Reference 15 is long and technical.
 

Attachments

  • HiFiNews-Jan78-p77-79.pdf
    230.3 KB · Views: 22
  • HiFiNews-Jun78-p81-85.pdf
    264.8 KB · Views: 12
hey-Hey!!!,
I suspect most of the issues I have with SS has to do with variable input capacitance based on voltage across the device. When I've taken trouble to neutralize this effect SS sounds good. I've also seen other attempts at neutralizing the variable input capacitance, with as near as I can tell the same results.
cheers,
Douglas
 
Some comments on Hawksford's article and two pages about traps in semiconductor materials.
 

Attachments

  • JAES1984-p542-543.pdf
    387.2 KB · Views: 13
  • traps.pdf
    155.2 KB · Views: 7
Back
Top Bottom