This is a good read. If tempted to use TL431 read the stability graphs. At the higher voltages they are less trouble. Read spec sheets as some have much better noise. For low current they work much like a zener. I have found the stablity is exactly as stated.
http://www.tnt-audio.com/clinica/regulators_noise2_e.html
I saw the rbb Va plot from AoE and suggested from the title on that plot that AoE was showing the answers to the formula.
"curves are simple theory: En² = 4kT(rbb' + re/2)"
here is what I stated, in that other Thread.
http://www.diyaudio.com/forums/analogue-source/154210-mpp-532.html#post4609730
It appears that instead, the plot is from noise measurements made by yourselves.
I wish someone had come forward sooner to confirm that the plots were based on noise measurements of the devices.
H & H, thank you for the very valuable noise measurements/reports.
Last edited:
Yes indeed. I researched and purchased the candidate transistors, Paul took the spectral measurements, and I processed the data, and created the spreadsheets and graphs. Unfortunately I didn't discover the ZTX1051 and ZTX1149, so we didn't get measurements or include them in our Table 8.1a, page 501. However, I think the ones we did find among the 60 or so candidates, had lower noise, although maybe also lower beta.
Thanks winhill2 for all the work finding these devices. bcarso alerted me to AoE 3rd and the ZTX851/951.
The ZTX851 do actually measure quite well in a test moving coil circuit I've been working on. I'm seeing results consistent with an rbb of less than 2 Ohms at 7.5 mA.
Pro Audio Design Forum • View topic - A Low Noise Balanced In Moving Coil Preamp Using the ZTX851
Balanced Input Flat Moving Coil Preamp for the Phono Transfer System
The results and an FFT with a 3R3 Rsource and a 1R Rgain are here and come in around -143 dBu Ein: Pro Audio Design Forum • View topic - A Low Noise Balanced In Moving Coil Preamp Using the ZTX851
The ZTX851 do actually measure quite well in a test moving coil circuit I've been working on. I'm seeing results consistent with an rbb of less than 2 Ohms at 7.5 mA.
Pro Audio Design Forum • View topic - A Low Noise Balanced In Moving Coil Preamp Using the ZTX851
Balanced Input Flat Moving Coil Preamp for the Phono Transfer System
The results and an FFT with a 3R3 Rsource and a 1R Rgain are here and come in around -143 dBu Ein: Pro Audio Design Forum • View topic - A Low Noise Balanced In Moving Coil Preamp Using the ZTX851
Last edited:
I read the original post that started this thread. I think if you are willing to go SMD, you might have a better chance to find good low noise transistors. You can even find more matching pairs. I can only speak for myself that I rather build with SMD than through hole circuit.
It might take a little getting use to, but it is a lot easier if you have the right equipment. A pair of very sharp pointed tweezers is a must to pickup the resistor and IC, you need two soldering iron to "pick" the component off the pcb. Everything is just smaller. I work with 0402, 0603, SOT-23 everyday. I take them over through hole components any time of the day. A good magnifying desk lamp and "Big Vision" glasses will help a great deal. I need to use both together to work with SOT23-6 opamps.
the biggest pit fall is it's hard to layout a two layers board using SMD, the pcb cost is higher with SMD.
JMHO
It might take a little getting use to, but it is a lot easier if you have the right equipment. A pair of very sharp pointed tweezers is a must to pickup the resistor and IC, you need two soldering iron to "pick" the component off the pcb. Everything is just smaller. I work with 0402, 0603, SOT-23 everyday. I take them over through hole components any time of the day. A good magnifying desk lamp and "Big Vision" glasses will help a great deal. I need to use both together to work with SOT23-6 opamps.
the biggest pit fall is it's hard to layout a two layers board using SMD, the pcb cost is higher with SMD.
JMHO
Last edited:
If only it were that easy.
I think most folks here, including myself, prefer something that they can actually work with and solder in place without the need for special glasses and other equipment.
In addition, some people's vision may prevent them from working with some of those microscopic devices.
I've used some 1206 resistors and caps without much issue, but anything smaller than that...forget it.
Hi Ammel68
I was a staunch opponent using SMD on DIY before. But since I was hired as a contractor, I have to do everything myself. I have to say I absolutely change my mind. I was with you not that long ago.
The solder iron is a little expensive, I use a Metcal. But the rest are cheap. The Big vision, desk magnifier and the tweezers total about $60.
i can do it so much faster. BTW, I am over 60, my eyes are pretty bad already. I even need the big vision on the through hole boards. After working with SMD for a while, it gets easy. Particular rework is so much faster. Stay away from 0402, even 0603 is not hard. But to be safe, use 0805, it has value on it, even if you drop it, you can still read the value.
As I said, the biggest thing is the pcb. It's hard to do it in 2 layers as it's hard to pass trace under the components. It's more expensive to do it in 4 layers. But I am seriously consider doing the IPS section in SMD. You can find match pairs much easier. You just have a lot more selections.
I was a staunch opponent using SMD on DIY before. But since I was hired as a contractor, I have to do everything myself. I have to say I absolutely change my mind. I was with you not that long ago.
The solder iron is a little expensive, I use a Metcal. But the rest are cheap. The Big vision, desk magnifier and the tweezers total about $60.
i can do it so much faster. BTW, I am over 60, my eyes are pretty bad already. I even need the big vision on the through hole boards. After working with SMD for a while, it gets easy. Particular rework is so much faster. Stay away from 0402, even 0603 is not hard. But to be safe, use 0805, it has value on it, even if you drop it, you can still read the value.
As I said, the biggest thing is the pcb. It's hard to do it in 2 layers as it's hard to pass trace under the components. It's more expensive to do it in 4 layers. But I am seriously consider doing the IPS section in SMD. You can find match pairs much easier. You just have a lot more selections.
Last edited:
Hi Ammel68
I am talking in the general sense, but you actually tickle my curiosity. I look around a little, here are some. But you better read in more detail to determine, don't take my words.
http://www.analog.com/media/en/technical-documentation/data-sheets/SSM2212.pdf
http://www.nxp.com/documents/data_sheet/BCM847BV_BS_DS.pdf
http://www.nxp.com/documents/data_sheet/PMP4201V_G_Y.pdf
Too many for me to list. Here is the search is use in Digikey:
https://www.digikey.com/product-search/en/discrete-semiconductor-products/transistors-bjt-arrays/1376378?k=transistor&pv74=158&pv69=3&FV=fff40015%2Cfff8007a%2C1280008%2C128009e%2C12800a3&mnonly=0&newproducts=0&ColumnSort=0&page=1&quantity=0&ptm=0&fid=0&pageSize=25
They tend to be lower voltage, but this is not an issue as people tend to put a cascode, you can use cascode to limit the voltage to below 40V.
Then you can find PNP for current mirror on the top. so you pretty much cover the IPS section. Even if you want complementary LTP as IPS, you can use the same matched NPN, PNP pairs.
Again, I am just give out some hot air. So take a look as see what you can find.
Last edited:
Also a dual matched jFET
https://www.digikey.com/product-detail/en/toshiba-semiconductor-and-storage/2SK2145-BL%28TE85L,F/2SK2145-BL%28TE85LFCT-ND/4304257
I notice the price is a lot cheaper also. I remember the MAT-01 etc. are over $20 ea.!!!
https://www.digikey.com/product-detail/en/toshiba-semiconductor-and-storage/2SK2145-BL%28TE85L,F/2SK2145-BL%28TE85LFCT-ND/4304257
I notice the price is a lot cheaper also. I remember the MAT-01 etc. are over $20 ea.!!!
Last edited:
Thanks winhill2 for all the work finding these devices. bcarso alerted me to AoE 3rd and the ZTX851/951.
The ZTX851 do actually measure quite well in a test moving coil circuit I've been working on. I'm seeing results consistent with an rbb of less than 2 Ohms at 7.5 mA.
Pro Audio Design Forum • View topic - A Low Noise Balanced In Moving Coil Preamp Using the ZTX851
Balanced Input Flat Moving Coil Preamp for the Phono Transfer System
The results and an FFT with a 3R3 Rsource and a 1R Rgain are here and come in around -143 dBu Ein: Pro Audio Design Forum • View topic - A Low Noise Balanced In Moving Coil Preamp Using the ZTX851
Thanks for the note, I appreciate that. It's interesting to note that in our measurements the PNP complement, ZTX951, had an even lower rbb', 1.24 ohms vs 1.67 ohms. You could easily refit your design to use PNP instead of NPN. The PNP parts measured with nearly as high beta as the NPN, see the beta plots in Figure 8.39, referenced from Table 8.1a.
Thanks for the replies.
I do want to try the ZTX951 PNP so that I can directly compare it to the 2SB737 samples I have. I have both the ZTX851 and ZTX951 in stock. I went with an NPN version first thinking that if Zetex tried to kill these parts off they would probably discontinue the PNP first. My hunch is that the NPN has highest volumes and would be the last of the two to go. I see that the earliest uses for these low-sat devices were in emergency fluorescent exit lights and car horns/sirens. The market's volumes for fluorescent exit lights surely must be falling due to LEDs so I hope these parts continue to find large volume applications - like low noise preamps 🙂.
WRT to emitter connections the bias is supplied by op amps A and B through the 316Ω resistors. The 1240 common mode rejection stages strip the common mode offset present at the op amp outputs. Current sinks could be placed from the emitters to V- but they really aren't necessary.
I have found that 2X e-line ZTX851/951 transistors fit neatly in an M7 nut. By epoxying the two devices inside the nut the temperature-induced offsets from air currents are greatly reduced. Though the output servos HPF that drift having the devices in close thermal contact reduces servo correction.
A pair of ZTX851 in an M7 nut to thermally couple them.
I do want to try the ZTX951 PNP so that I can directly compare it to the 2SB737 samples I have. I have both the ZTX851 and ZTX951 in stock. I went with an NPN version first thinking that if Zetex tried to kill these parts off they would probably discontinue the PNP first. My hunch is that the NPN has highest volumes and would be the last of the two to go. I see that the earliest uses for these low-sat devices were in emergency fluorescent exit lights and car horns/sirens. The market's volumes for fluorescent exit lights surely must be falling due to LEDs so I hope these parts continue to find large volume applications - like low noise preamps 🙂.
WRT to emitter connections the bias is supplied by op amps A and B through the 316Ω resistors. The 1240 common mode rejection stages strip the common mode offset present at the op amp outputs. Current sinks could be placed from the emitters to V- but they really aren't necessary.
I have found that 2X e-line ZTX851/951 transistors fit neatly in an M7 nut. By epoxying the two devices inside the nut the temperature-induced offsets from air currents are greatly reduced. Though the output servos HPF that drift having the devices in close thermal contact reduces servo correction.
A pair of ZTX851 in an M7 nut to thermally couple them.
Last edited:
Hello
I have a lot ultra low nois Transistors build in the Accuphase C280 MC Stage an in the big external MC amp. 2SA 1114 and 2SC2602
Regards
I have a lot ultra low nois Transistors build in the Accuphase C280 MC Stage an in the big external MC amp. 2SA 1114 and 2SC2602
Regards
Found this paper from 1980 that shows an NPN transistor specifically optimized for moving coil phono cartridge amplifiers (!). Rbb = 1.4 ohms, only slightly less than ArtOfElectronics's measurement on the ZTX851 discrete NPN (1.7 ohms). Interesting that they took into account the nonuniform current density in the emitter metallization.
I have 2 diff pairs and 2 fairly mundane integrated pre-amps (never published) that the authors sent as a professional courtesy. These are probably by now the only ones left in existence. The pre-amps are hard wired at G=50 with 7 - 50 Ohm base resistors in series and 7 in parallel as a feedback network. By design it needs a huge electrolytic to ground on the gain resistor.
Of course the diff pairs have the emitters tied (why?) so I could never make a pair of in-amps out of them.
Last edited:
Thanks for sharing that paper with us.
I also found this base topology in a Zetex application note:
Zetex DN-11 shows the ZTX650/690/750/790 in a low noise MC preamp http://www.proaudiodesignforum.com/images/pdf/Zetex_DN11_Low_Noise_Audio_Input_Stage.pdf
I also found this base topology in a Zetex application note:
Zetex DN-11 shows the ZTX650/690/750/790 in a low noise MC preamp http://www.proaudiodesignforum.com/images/pdf/Zetex_DN11_Low_Noise_Audio_Input_Stage.pdf
Disk drive head amps used a similar trick. I have a die photo of an SK170 somewhere with the same type of arrangement to maximize Z/L with area.