|
Post by pandapops on Jul 21, 2012 21:33:20 GMT
Lots of amps have these as their first stage, presumebly they have a disproportionate effect on the end result, yet I don't see people rolling transistors even though small signal transistors are cheap unlike opamps and tubes.
Are there just too many variables involved? What parameters would need to similar before an experiment is sensible to even attempt?
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 21, 2012 21:48:46 GMT
For more info on this look in the books from Douglas Self for instance.
Subjectivism is lurking behind every leg of each component... If you do experiment build 2 of each with only one thing changed and compare blind if possible. It will bring (misguided ?) wisdom if you do. Also will the way of testing determine if you are searching for a definite or an indefinite period of time.
|
|
|
Post by pandapops on Jul 21, 2012 22:13:19 GMT
Transistor selection would be guided by datasheets, final assessment subjective.
An example is an old amp I have that starts with a LTP of BC182, the datasheet for these has a noise figure of 10db which is not good by modern standards, however will there be other parameters rendering it a decent trade off?
Another example is the Panda that has jfets, my understanding is that they are low noise but will output some dc.
With these cheap bits performing a simple task, isn't it possible to approach a best possible solution?
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 21, 2012 22:49:38 GMT
Lots of amps have these as their first stage, presumebly they have a disproportionate effect on the end result, yet I don't see people rolling transistors even though small signal transistors are cheap unlike opamps and tubes. Are there just too many variables involved? What parameters would need to similar before an experiment is sensible to even attempt? Pandapops Ignore smart ass comments like that from Frans. He and his fellow EEs Douglas Self, and Rod Elliot for example, have closed minds as far as the subjective effects of different types of components and cables etc. as long as they measure approximately the same. Like most of his ilk, he doesn't have a bloody clue about the limitations of human hearing either. He and his fellow EEs still think mainly in terms of sine wave response of the human ear, when it can be shown these days that the rise time of a waveform is also important. Douglas Self does however show in his various books , the greatly reduced efects on distortion with close matching of the LTP to within 1%, and indeed recommends this. Just ignore the acidic comments about subjectivity that Douglas Self makes in the preface of his books and you will find they will give a lot of good information that can often be used to further improve the technical specifications of an amplifier. The Input pair devices may be replaced by electrically similar but lower noise devices matched by both HFE and VBE. In the case of the Class A amplifier preamp and amplifier projects in R.G, we are doing this by using dual metal can devices that are very closely matched by the manufacturer. Quite a few of us have however had good results by thermally bonding together selected devices chosen by measuring HFE and VBE to close tolerances. The price of some devices such as the BC550 and BC560 as an example, when purchased in 100 or more quantities makes it relatively easy to obtain quite a few matched devices when they come from the same manufacturing batch. Regards Alex P.S. Input FETs are much harder to match electrically.
|
|
|
Post by pandapops on Jul 22, 2012 0:26:43 GMT
Thanks Sandy, I hadn't considered the vbe values.
Are there affordable double transistors then? I was under the impression that these were falling out of use and are seriously expensive, I've seen some at £17 each.
Is multimeter matching adequate, I've been told that proper HFE matching should be 5v at a specified current and frequency, with the HFE having an inverted U curve as the current increases. This was by an electronics supplier after I complained about some transistors testing well below spec on a multimeter.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 22, 2012 0:42:47 GMT
Pandapops You can buy the Motorola NPN 2N2920A and the PNP 2N3811A dual transistors for around AU$10 each on ebay. We are presently organising a GB for the LS352 PNP dual and the LS313 dual for the Class A amplifier.Check out the data sheets for those mentioned. The "A" versions of the 2N3811 and 2N2920 are much more closely matched for HFE and VBE. The LS352 from Linear Devices is a modern replacement for the 2N3811A. You really need more than just a DMM to select matched devices. Several members have commercial transistor testers for this purpose. Regards Alex
|
|
Will
Been here a while!
Ribena abuser!
Member since 2008
Posts: 2,164
|
Post by Will on Jul 22, 2012 9:46:02 GMT
Hi Pandapops, The devices that Alex (Sandyk) mentions that we are buying are: LS313 & LS353They are basically two transistors in the same package, very tightly matched for for hFE and vbe. The same company also does similar matched jfets, such that would be of use in the LTP input of your panda. If you want some of the LS313/352, please PM me. The transistor tester a few of us have got is the Atlas DCA55. Superb bit of kit, which paid for itself pretty much straight away.
|
|