Motorola marks the 200Vce MJE341/MJE344 at 15pF Cob.
Whether Cob of the 340/350 is 25pF, or 20 for real, makes no difference for the choice to make.
Mr Cordell measured fT of the 340/350 at 18MHz for the NPN, and 30MHz for the PNP (is no typo)
From here :
* MJE340C - created March 7, 2011
.MODEL MJE340C npn
+IS=800e-15 BF=180 VAF=100
+IKF=0.35 ISE=25e-12 NE=1.5
+RB=21 RC=2 RE=0.01
+CJE=170e-12
+CJC=140e-12
+TF=7600e-12 XTF=10 VTF=10 ITF=1
+TR=10000e-12 BR=0.004 IKR=0.05
+EG=0.64 NC=2
+ISC=1.5e-10 VAR=100
This is the spice model for the said part. The manufacturer does physical measurement of the node capacitances.
See CJC which is the base collector capacitance measured at 0V.
Ive seen even worse measurement of this as there are manufacturing differences between companies which account for it.
Is it not rather 14e-12 , that is 14pF.?...
Motorola s model from 2004 say 10pF , quite large variations...
Some manufacturers measure at 10 V and some at 5 V and then some add 30 percent for the value at those voltages at 0 V, confusing stuff but probably done to improve overall accuracy of the model.
That model is Bob Cordells model, probably derived from the fairchild or Onsemi models.
My old siemans model has it as 15.6 pf.
In any case this is far to high for vas use, and as said FT is low as well which worsens matters.
That model is Bob Cordells model, probably derived from the fairchild or Onsemi models.
My old siemans model has it as 15.6 pf.
In any case this is far to high for vas use, and as said FT is low as well which worsens matters.
I think we know that they would not be subjected to 100V, but rather les than the rail voltage ( in this case, less than 50V ), the output FET's do require some of that rail voltage to turn on, but spike's/accidents do happen that is why it is always recommended to choose devices that can cope to at least the rail to rail voltage.
As for the max possible power on 8 Ohm load with 50V rail and Fet output devices, well it can actually go higher than 100W, the fets will take less than 10V to turn on so the power can go up to 110W or even a bit more.
As for the max possible power on 8 Ohm load with 50V rail and Fet output devices, well it can actually go higher than 100W, the fets will take less than 10V to turn on so the power can go up to 110W or even a bit more.
Bd139 could probably work as the voltage accross it would be below 100 v but once again its a poor choice. It has too high cob about 12-14 pf, low early voltage. In this case both high and low frequency distortion gets affected. If these better parts are hard to find its better to use a common transistor like mpsa92 and paralell 2 of them. Typical cob for 1 is 1.5-2 pf so 2 in paralell for a max of 4 is pretty good and gives performance similar and depending on the parts used could even better the performance of the purpose designed video transistors. I myself use bf721 parts in paralell many times. A small 1 or 2 ohm resistor on the emitters is enough for matching when there is a 10 percent match between two parts.
I can read graphs, don't have to be Typical values.
When a manufacturer mentions a Typical value of 6pF, while the graph reads 2, it appears to me that the parameter spread is generous.
( http://www.diyaudio.com/forums/soli...a-24-quad-received-last-nite.html#post2745901 )
When a manufacturer mentions a Typical value of 6pF, while the graph reads 2, it appears to me that the parameter spread is generous.
( http://www.diyaudio.com/forums/soli...a-24-quad-received-last-nite.html#post2745901 )
3 transistors for VAS? then why not 4 or 5 for that matter...
I would have a question for all of you who seek the so called hi-fi sound: The hy freq distorsions that you acused MJE340 of givin due to it's Cob ( wich it's taken from some spice model in stead of it's datasheet ), are you able to hear that distortions? and i mean with your own ears and not testing it on some fancy device? cus i think it's just a waste of time and resources to solve a problem that in fact does not exists, i doubt verry much that there are too many people that can make the difference between 0,02 and 0,002 THD, you cant because it is too low for the human ear to pick it up, so why do the big companyes waste time and money on this subject? It's simple, it is a very efficient way to justify huge pricesc for theyr amplifiers, no one cares that they will not be able to listen to those minuscul harmonics, no one cares that they more often cannot tell the difference, they have an hi-fi amplifier that const thousends of dollars, that is all that matters right? maybe not...
Anyway, it's just a humble opinion by someone who studdied the marketting strattegies enough to tell, so don't shoot me
Mpsw 42/92 may be an option as HM mentions but Bf87x can still be sourced in repair or DIY volumes.
In most well designed circuits, the use of MJE340/350 vs lower Cob transisitors in the VAS, can easily be heard. HF detail is masked in the former. Attack and Slam being that much lesser is obvious with Complex Music. Ofcourse, the rest of your equipment chain should also be high resolution.
No shooting or flames......I would have a question for all of you who seek the so called hi-fi sound: The hy freq distorsions that you acused MJE340 of givin due to it's Cob ( wich it's taken from some spice model in stead of it's datasheet ), are you able to hear that distortions? and i mean with your own ears and not testing it on some fancy device? so don't shoot me
As far as I know, these On-semi parts have been used in countless VAS and driver stages but simulate and test with higher distortion than when typical video amplifiers are used as the VA. Ignoring the technical basis widely expressed here, audiophiles will generally hold a preference for the fastest, lightest VAS on subjective grounds. It's unusual to find both audiophile and technical views in agreement like this.
Here in Oz, we have used video drivers for ~40 years in DIY - plenty of experience in that period and subjectively, using the On-semi types for any reason usually resulted in some degree of degraded sound which varies from the psycho-acoustic stereo effects to a simple sense that treble detail is less clear etc. I built several SC magazine designs specifying them ~1990, ranging from 50-200W and then improved the sound subjectively by simply substituting readily available video driver types BF469/70. The difference may not have been instantly obvious but after a few hours listening, many were a lot happier. I have also repaired several commercial amps that used obsolete and slow parts (even for PA) by substituting video drivers and had client comments in agreement, so this is not just my opinion.
Interestingly, I see a number of designs from eastern Europe on the net that use the On-semi parts so I guess they became popular in your region, perhaps also for availability or cost reasons.
Last edited:
I'm more of a pragmatic one so my opinion might be biased, i always desighn my own amps ( or most of anythyng else i built for that matter ) and i base my desighns mostly on cost effectiveness but not sacrificing quality, i do not seek the perfect audio amplifier ( nor do i think it actually exists ), hi-fi is most of the time for people with big pockets and i am not one of them, this is one of my latest desighns:
Anyway, as many people say, the audio domain is one of the most affected by subjectivity, each has it's own opinion, i do not seek to upset anyone nor do i want to impose my way of viewing things, i only express my opinions as i see it and hope nowone takes offence by it
Anyway, as many people say, the audio domain is one of the most affected by subjectivity, each has it's own opinion, i do not seek to upset anyone nor do i want to impose my way of viewing things, i only express my opinions as i see it and hope nowone takes offence by it
Attachments
Last edited:
I strongly recommend that you try one of the more modern designs using faster semiconductors. Examples: Mike's Symassym (updated version), Lazy Cat's SSA, TSSA or even Quasi's N-channel in which case, I'd say substitute the MJEs with Video Transistors and hear for yourself what everyone else is talking about here.
3 transistors for VAS? then why not 4 or 5 for that matter...
I would have a question for all of you who seek the so called hi-fi sound: The hy freq distorsions that you acused MJE340 of givin due to it's Cob ( wich it's taken from some spice model in stead of it's datasheet ), are you able to hear that distortions? and i mean with your own ears and not testing it on some fancy device? cus i think it's just a waste of time and resources to solve a problem that in fact does not exists, i doubt verry much that there are too many people that can make the difference between 0,02 and 0,002 THD, you cant because it is too low for the human ear to pick it up, so why do the big companyes waste time and money on this subject? It's simple, it is a very efficient way to justify huge pricesc for theyr amplifiers, no one cares that they will not be able to listen to those minuscul harmonics, no one cares that they more often cannot tell the difference, they have an hi-fi amplifier that const thousends of dollars, that is all that matters right? maybe not...
Anyway, it's just a humble opinion by someone who studdied the marketting strattegies enough to tell, so don't shoot me
If 3 transistors provide the lowest Cob with SOA needed I would but there are parts that perform better as single parts from 2 in parralell on. Krell use 7 of these transistors to make a super driver btw, consult some of their schematics.
The same method is used to obtain some of the datasheet performance figures that is used to obtain the spice models. How do you suppose the device capacitances are obtained.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.