Re: More pedigreed audio tansistors
Harry:
followed up your list (filtered by availability at Reichelt):
SA1209/2911 makes a very nice predriver or voltage gain pair, the nicest I have seen in TO126 case
The SA1085E seems to be an extraordinarily low noise PNP. Unfortunately, there is no complementary. But then I sometimes wonder on what basis manufactures recommend complementaries... Early voltage is always different (has to be), h_FE curve is also often different with older devices..
Eric
HarryHaller said:
Harry:
followed up your list (filtered by availability at Reichelt):
SA1209/2911 makes a very nice predriver or voltage gain pair, the nicest I have seen in TO126 case
The SA1085E seems to be an extraordinarily low noise PNP. Unfortunately, there is no complementary. But then I sometimes wonder on what basis manufactures recommend complementaries... Early voltage is always different (has to be), h_FE curve is also often different with older devices..
Eric
better yet: 2SA1538/SC3953
http://www.nedis.com/datasheets/datasemi/2SC/2SC3953.pdf
seriously the best drivers I have seen so far: f_T = 400 MHz, cob around 2 pF, I_C(V_CE) also looks too good to be true. Even better than SA1404-1407!
TO-126 isolated case.
Drawback: Sanyo no longer lists them but they are still available at numerous distributors.
Eric
http://www.nedis.com/datasheets/datasemi/2SC/2SC3953.pdf
seriously the best drivers I have seen so far: f_T = 400 MHz, cob around 2 pF, I_C(V_CE) also looks too good to be true. Even better than SA1404-1407!
TO-126 isolated case.
Drawback: Sanyo no longer lists them but they are still available at numerous distributors.
Eric
Data Sheets
Some (not all) Toshiba data sheets
http://www.semicon.toshiba.co.jp/eng/search/ft_datasheet.html
H.H.
Some (not all) Toshiba data sheets
http://www.semicon.toshiba.co.jp/eng/search/ft_datasheet.html
H.H.
Years ago in "Audio Conversions" magazine, Les Sage described an Active Loudspeaker System using discrete unity gain buffers, he made a big deal about the bipolars he choose, Philips <a href="http://www.semiconductors.philips.com/pip/BFQ232.html">BFQ232</a> and <a href="http://www.semiconductors.philips.com/pip/BFQ252.html">BFQ252</a>, he specified the BFQ232<b>S</b>/252<b>S</b> but they don't appear on the Philips site, data sheets are a bit minimal 
.
<hr width="95%" align=center>From the Article "The Devices chosen are in fact medium power, ultra linear, ultra high bandwidth, CRT drivers'
<table width="65%" border="1" cellspacing="0" cellpadding="1" >
<tr>
<td ><center>---</center></td>
<td ><center>V<sub><font size="-1">CBO</font></sub></center></td>
<td ><center>I<sub><font size="-1">C</font></sub> DC</center></td>
<td ><center>H<sub><font size="-1">fe</font></sub></center></td>
<td ><center>F<sub><font size="-1">T</font></sub></center></td>
</center></tr>
<tr>
<td ><center>BFQ232/252</center></td>
<td ><center>100/-100 V</center></td>
<td ><center>300/-300 mA</center></td>
<td ><center>30 typ.</center></td>
<td ><center>1.4 / 1.3 GHz</center></td>
</tr></table>
The closest alternative devices he could find were the 2SA1404D/2SC3598D, but he did consider them second best.
The BFQs now appears to be discontinued, but still readily available.
Regards
James
.<hr width="95%" align=center>From the Article "The Devices chosen are in fact medium power, ultra linear, ultra high bandwidth, CRT drivers'
<table width="65%" border="1" cellspacing="0" cellpadding="1" >
<tr>
<td ><center>---</center></td>
<td ><center>V<sub><font size="-1">CBO</font></sub></center></td>
<td ><center>I<sub><font size="-1">C</font></sub> DC</center></td>
<td ><center>H<sub><font size="-1">fe</font></sub></center></td>
<td ><center>F<sub><font size="-1">T</font></sub></center></td>
</center></tr>
<tr>
<td ><center>BFQ232/252</center></td>
<td ><center>100/-100 V</center></td>
<td ><center>300/-300 mA</center></td>
<td ><center>30 typ.</center></td>
<td ><center>1.4 / 1.3 GHz</center></td>
</tr></table>
The closest alternative devices he could find were the 2SA1404D/2SC3598D, but he did consider them second best.
The BFQs now appears to be discontinued, but still readily available.
Regards
James
Philips Transistors Spice
Thanks James. I stumbled on some Spice models thanks to your link.
http://www-us10.semiconductors.philips.com/models/spicespar/
H.H.
Thanks James. I stumbled on some Spice models thanks to your link.
http://www-us10.semiconductors.philips.com/models/spicespar/
H.H.
BFQ devices and Zetex models
The BFQ devices have fantastic bandwidth and low collector-base capacitance. However, current gain is low and very non-linear. Wouldn't use them except in darlington configuration. Sanyo makes much nicer video transistors to be used as audio drivers, though they ain't cheap, let me tell you (just ordered 180€ worth of transistors...)
Harry,
the Zetex pages really suck in terms of organisation and search capabilities. I only stumbled on the master model file after I had used the site many times.
Eric
The BFQ devices have fantastic bandwidth and low collector-base capacitance. However, current gain is low and very non-linear. Wouldn't use them except in darlington configuration. Sanyo makes much nicer video transistors to be used as audio drivers, though they ain't cheap, let me tell you (just ordered 180€ worth of transistors...)
Harry,
the Zetex pages really suck in terms of organisation and search capabilities. I only stumbled on the master model file after I had used the site many times.
Eric
Re: Data sheets
Count me in! Also Power BJT/MOSFET/IGBT.
Eric is right about Zetex Website. You have take your time to find the things you need.
I can see Sanyo produce some nice transistors.. Funny i have not been looking at them before. Check also Rohm but they have not Spice models at all.. I could not find them.
Sonny
HarryHaller said:
Count me in! Also Power BJT/MOSFET/IGBT.
Eric is right about Zetex Website. You have take your time to find the things you need.
I can see Sanyo produce some nice transistors.. Funny i have not been looking at them before. Check also Rohm but they have not Spice models at all.. I could not find them.
Sonny
Geoff was so kind as to e-mail me the SA1349/SC3381 data sheets. That Taiwanese server still does not like me. Are there any other interesting sheets on that server?
The Toshiba duals look pretty good, though they are not as linear in h_Fe as the best single Toshiba transistors or the MAT02/03 aka SSM2210/20 (now obsolete, sadly). Capacitances are better than MAT02/03 though.
Toshiba write "reccommended for caascode, current mirror circuit applications of the first stages of pre, main amplifiers". Do they mean they should be used with cascode and mirror or as cascode and mirror components? In my eyes, the largest benefit can be had from using them as input pairs. If one feels like littering the circuit with 90 €cent / piece parts, that can be done, but the added benefit would be negligable, especially if the current mirror has emitter degeneration resistors (the cascode transistors only need to have some matching of h_FE anyway).
The substrate lead is to be left open. I wonder if you can further improve the performance by tying the substrate to a well-decoupled supply or bootstrapping it. AC- bootstrapping it to the output devided by the global amplification could remove the miller capacitance.
Sonny,
I find Sanken LAPT such as SA1186/SC2837 unmatched as output bipolars in terms of gain linearity and bandwidth even though they have been around for about 15 years. Do you (or anybody) know about better output devices? What is the situation on SPICE models for Sanken?
Greetings,
Eric
Eric
The Toshiba duals look pretty good, though they are not as linear in h_Fe as the best single Toshiba transistors or the MAT02/03 aka SSM2210/20 (now obsolete, sadly). Capacitances are better than MAT02/03 though.
Toshiba write "reccommended for caascode, current mirror circuit applications of the first stages of pre, main amplifiers". Do they mean they should be used with cascode and mirror or as cascode and mirror components? In my eyes, the largest benefit can be had from using them as input pairs. If one feels like littering the circuit with 90 €cent / piece parts, that can be done, but the added benefit would be negligable, especially if the current mirror has emitter degeneration resistors (the cascode transistors only need to have some matching of h_FE anyway).
The substrate lead is to be left open. I wonder if you can further improve the performance by tying the substrate to a well-decoupled supply or bootstrapping it. AC- bootstrapping it to the output devided by the global amplification could remove the miller capacitance.
Sonny,
I find Sanken LAPT such as SA1186/SC2837 unmatched as output bipolars in terms of gain linearity and bandwidth even though they have been around for about 15 years. Do you (or anybody) know about better output devices? What is the situation on SPICE models for Sanken?
Greetings,
Eric
Eric
Web site organization
I agree...... It takes quite a bit of wandering about to find stuff in Semiconductor websites but I guess I am just nosey. It is not as quite as bad as playing Adventure! on a Vax 750 at work. Any of you fellow seasoned citizens remember that pre-graphics computer game?
"You are in a maze of twisty little passages, all alike."
H.H.
I agree...... It takes quite a bit of wandering about to find stuff in Semiconductor websites but I guess I am just nosey. It is not as quite as bad as playing Adventure! on a Vax 750 at work. Any of you fellow seasoned citizens remember that pre-graphics computer game?
"You are in a maze of twisty little passages, all alike."
H.H.
Toshiba duals substrate connection
Had another look at MAT02/03 data sheets and spice models: There are parasitic diodes from each emitter and collector to the substrate that are contained in those manufacturer models.
Reverse biasing (i.e. tying the substrate to the most negative point for NPN and vice versa for PNP) will reduce the capacitance (see typical plot). In a differential pair or current mirror configuration, the capacitive coupling of the collectors will give a (non-linear) reduction in bandwidth. AC-grounding the substrate would eliminate that path.
So I think against Toshiba's recommendation, the substrate lead should not be left open. I still think bootstrapping would also be attractive but attaining stability in a closed loop configuration might become a bear...
Eric
Had another look at MAT02/03 data sheets and spice models: There are parasitic diodes from each emitter and collector to the substrate that are contained in those manufacturer models.
Reverse biasing (i.e. tying the substrate to the most negative point for NPN and vice versa for PNP) will reduce the capacitance (see typical plot). In a differential pair or current mirror configuration, the capacitive coupling of the collectors will give a (non-linear) reduction in bandwidth. AC-grounding the substrate would eliminate that path.
So I think against Toshiba's recommendation, the substrate lead should not be left open. I still think bootstrapping would also be attractive but attaining stability in a closed loop configuration might become a bear...
Eric
capslock said:
Bootstrapping the substrate would definently be worth trying out!! But i would suggest a measurement on their behavior. AC_grounding would improve their performance regarding capacitive coupling of noise.
hhhrrrmmpphh (hehehe) I have bouild Three different amps using 2SA1186 and 2SC2837...
One with a short path and inverting mode.. Took me 3 days to make from the scratch and is one the best sounding amp i have heard (I have only heard something like the small Rotels .. But is also a cheap speaker!) with speakers like SystemAudio (Light loads). Did not produce more than 25W/8Ohm. I think my little brother have it now.
And one in Class A and no FeedBack loop. -- Sounded bad .. Not to blame on the Outputs.. My first real AMP.
And a CFB Wich run today with 2SA1302/SC3281. Sounds really nice
With GT20D101 and D201 the sound was a bit dark .. Lack warmth and crisp sound which the 2SA1186/SC2837 and the 2SA1302/SC3281 have. The reason to choose the last one is because of the power and i use only one pair.It was a hard time to get it up and working!!
What i am going to say ... It is a good pair like the bigger ones (2SA1302/SC3281) .. But i have not made comparison on their sound with two identical systems and is around 7 years ago i made the first one.
For the Spice model ... I have not other than the one from the one Harry has which somebody chewed his but for!?
Sonny
Just wanted to let you know that I posted a separate thread on the 2SA1349 dual transistors in the "electronics and parts" forum. Unwittingly, this evolved into a discussion on good and bad second sources on power amp output transistors and in a comparison between Toshiba, Moterola and Sanken parts which would have belonged here.
Check it out!
Eric
Check it out!
Eric
Need Your Help About Transistors
Hi Jacko,
You seemed well versed with transistors, so maybe you could help me. My amplifier is using a 2SC2362 type of transistor. Somebody suggested to me that by replacing this with a 2SC2240GR will further improve its sonic performance. Does it means it is a direct replacement? Also, how do I get a match pair ( using a multimeter? )?. How do I get the highest gain possible? I'm just getting a second opinion.
Thanks in advance.
Edwin
Hi Jacko,
You seemed well versed with transistors, so maybe you could help me. My amplifier is using a 2SC2362 type of transistor. Somebody suggested to me that by replacing this with a 2SC2240GR will further improve its sonic performance. Does it means it is a direct replacement? Also, how do I get a match pair ( using a multimeter? )?. How do I get the highest gain possible? I'm just getting a second opinion.
Thanks in advance.
Edwin
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.