Nanaimo
A good substitution needs more than part numbers and their datasheets to decide whether they will perform similarly in an application. Generalised descriptions like "audio driver transistor" and "low level amplifier" used for these can apply to many, even very dissimilar transistor types so you need to specify the application circuit, voltage and signal levels at least.
In general though, we know that these are NPN, low voltage, low level amplifiers with reasonably low noise qualities too. KSC1845 is a high voltage, very low noise version for similar duties in elevated voltage circuits. It should work fine in the same applications at lower voltages too, but its ratings won't be needed. Wherever cost or availability is an issue, other parts like BC550 or 2SC1815 could be even better as a differential input (LTP) stage for example, as long as the voltage conditions are within spec.
Note that the 945,458 transistors are no longer the same as originals made by manufacturers NEC and Hitachi decades ago. What you buy could be any of the many similar types with variations like 2SC945A and in various Hfe ranges, different packages with odd suffix codes, still around and available from several sources. They are also marketed under several aliases but datasheets don't show all of them. Others parts may be just generic "no-name" rubbish. Make sure your comparisons are based on known quality rather than what Ebay sellers want you to believe.
NEC - datasheet pdf
Hitachi Semiconductor - datasheet pdf
HTTP 301 This page has been moved
In general though, we know that these are NPN, low voltage, low level amplifiers with reasonably low noise qualities too. KSC1845 is a high voltage, very low noise version for similar duties in elevated voltage circuits. It should work fine in the same applications at lower voltages too, but its ratings won't be needed. Wherever cost or availability is an issue, other parts like BC550 or 2SC1815 could be even better as a differential input (LTP) stage for example, as long as the voltage conditions are within spec.
Note that the 945,458 transistors are no longer the same as originals made by manufacturers NEC and Hitachi decades ago. What you buy could be any of the many similar types with variations like 2SC945A and in various Hfe ranges, different packages with odd suffix codes, still around and available from several sources. They are also marketed under several aliases but datasheets don't show all of them. Others parts may be just generic "no-name" rubbish. Make sure your comparisons are based on known quality rather than what Ebay sellers want you to believe.
NEC - datasheet pdf
Hitachi Semiconductor - datasheet pdf
HTTP 301 This page has been moved
Thanks for these answers! Having just returned from an Internet-free part in the County of North Norfolk UK) I must identify the application: Reel-to-Reel AKAI GX-635D tape recorder. Right Record channel dead. Left OK. Both L & R channels OK in replay. I have the User Manual with Cct. Diags and voltages shewn thereon next to the transistors. Note: the PCB shows a white rectangle around each transistor site. This rectangle is 'polarised' with the white thick end at pin 3. I have removed all the old input 2SC458 transistors from the PCB - ten (R) and ten (L) - and am now in the position to replace them all with these KSC1845's. I have a PEAK DCA55 meter which measures the Hfe of an out-of-circuit transistor and automatically gives each pin its number and whether it is B,C or E. I measured the Hfe of the old 2SC458's as I removed them. Ought I to use the replacement KSC1845's with, as near as possible, the same Hfe as each other and put them in regardless of the Hfe of the old 2SC458's? The input Cct is built more-or-less in 'daisy-chain' with no matching pairs. But of course the (L) and (R) channels each have their TR1 through to TR10 and all, with the exception of the two TR9's which were 2SC458's. TR9's, according to the Cct. Diag. should have been 2SC458's asd well but were, in fact BC109's. So they didn't keep to the specs anyway. When you say 'ratings' are you referring to this Hfe?
It's unlikely an Akai product was ever fitted with BC109 transistors during manufacture. These may have been fitted in an early attempt to do as you may be - trying to lower noise levels. I guess it's too late now but if you were to match the gain of these parts to those removed, it might have been helpful but again, it's unlikely any Hfe matching would have taken place then unless a long-tailed pair or similar balanced requirement indicated matching.
As long as Hfe is not greatly different to the values you have removed, it should be fine. If not, you could get instability or otherwise insufficient levels and poor sound quality.
Transistor ratings is a term usually meaning principle maximum DC + signal operating conditions such as specific voltages, current, power levels, dissipation/temperature etc. as shown in the first block of the relevant product data sheets. eg: http://www.farnell.com/datasheets/296634.pdf
As long as Hfe is not greatly different to the values you have removed, it should be fine. If not, you could get instability or otherwise insufficient levels and poor sound quality.
Transistor ratings is a term usually meaning principle maximum DC + signal operating conditions such as specific voltages, current, power levels, dissipation/temperature etc. as shown in the first block of the relevant product data sheets. eg: http://www.farnell.com/datasheets/296634.pdf
I should have explained Hfe grades in the above post. Each relevant transistor is Hfe graded on the schematic. G,D,P etc. suffixes are used to denote the specified Hfe grade or range of Hfe for each location. See the datasheet links in #2 for the details applying to each type. Note that you can no longer buy these many varieties - legacy or generic products are only available in 1 or 2 grades or unmarked. Sorting is the only practical option unless it is a non-critical application.
As you are using the shotgun approach to repairs, don't forget that capacitors also fail and much more likely with the passage of time. There are dozens of small value electrolytics in there, isolating DC from the signal at every stage in the preamplifiers and control circuits. Replacing these with good quality caps will be a worthwhile exercise whether transistors are shot or not.
As you are using the shotgun approach to repairs, don't forget that capacitors also fail and much more likely with the passage of time. There are dozens of small value electrolytics in there, isolating DC from the signal at every stage in the preamplifiers and control circuits. Replacing these with good quality caps will be a worthwhile exercise whether transistors are shot or not.
Thanks for this reply, Ian. I have downloaded the 2SC458 and the 2SC945 Data sheets. Is there any reason to suspect the configuration E.C.B.=1.2.3 on the original 2SC458 is any different from that shewn on the Data sheets? This splendid little Transistor checker I have (PEAK DCA55) shows the configuration of any transistor applied to it for analysis. I got a little worried there may have been a change - after I removed the twenty transistors. Sorry about the 'shotgun' approach I have adopted! I am trying NOT to lose sight of my reason to attempt a repair: (R) channel dead in Record Mode. All OK in Replay Mode. The noise level hadn't entered in the equation (yet!)
With the DCA55 in hand I catalogued the Hfe readings from each transistor as I removed them. Each transistor I then mounted on a small annotated card and splayed their legs to keep them in position.
TR9(L) and (R) - those BC109's. They were already changed when I bought the machine NEW in 1983. There were other transistors not complying with the List on P.59 on the Parts List for the Pre-Amp PC Board (TH-5001A) (This number is also stamped on the PCB itself.)
If you have this list everything will fall into place as I now make a list of the removed transistors along with their Name, position in Cct, and Hfe:
Transistor TRNo. L or R Hfe Ic Vbe Test Ib
C458 TR1 L 75 2.5mA 0.76v 4.64mA
C458 TR1 R 274 2.5mA 0.76v 4.64mA
C458 TR2 L 262 2.5mA 0.76v 4.64mA
C458 TR2 R 276 2.5mA 0.76v 4.64mA
C458 TR3 L 80 2.5mA 0.76v 4.64mA
C458 TR3 R 339 2.5mA 0.76v 4.64mA
C458 TR4 L 248 2.5mA 0.76v 4.64mA
C458 TR4 R 271 2.5mA 0.76v 4.64mA
C458 TR5 L 253 2.5mA 0.74v 4.64mA *
C458 TR5 R 281 2.5mA 0.76v 4.64mA *
C458 TR6 L 277 2.5mA 0.76v 4.64mA *
C458 TR6 R 257 2.5mA 0.76v 4.64mA *
C458 TR7 L 263 2.5mA 0.76v 4.64mA
2N2712 TR7 R 181 2.5mA 0.70v 4.70mA *
C458 TR8 L 301 2.5mA 0.76v 4.64mA
C458 TR8 R 313 2.5mA 0.76v 4.64mA
BC109 TR9 L 644 2.5mA 0.74v 4.66mA *
BC109 TR9 R 240 2.5mA 0.74v 4.70mA *
C458 TR10 L 142 2.5mA 0.76v 4.64mA *
C458 TR10 R 259 2.5mA 0.76v 4.64mA *
So this is what I removed. The * denote changes from original specs. as follows:
TR5,TR6, and TR10 (L&R) were spec'd as 2SC945L
TR9 (L&R) were spec'd as 2SC458LG(D) but were replaced with RSBC109's.
The PEAK DCA55 test unit uses 12v. battery and PEAK have made a demonstration video on YouTube. There's also a Capacitor tester - but I don't have one (yet!) In view of the comment about the caps in the PCB should I push the boat out a bit more and get one?? Michael (UK)
Hi
With electrolytic capacitors, because they do all dry out and corrode to useless scrap over time, the shotgun approach is indeed what you do. Just replace them all after listing from the parts list and buying the values needed. After so many years, the available values/ratings may be different now but there is no problem using higher voltage ratings or up to 2x the capacitance values in most low-level audio circuits. The important issue is to do it - there are too many to consider individually. You won't need a capacitance or ESR meter to prove anything when they are pulled - simply bin and replace.
I don't know what to say about the BC109s - they are plain vanilla, ungraded types and at odds with the original specs. - Id say it was an unplanned substitution whoever did it. BC109s are European (prolectron) types with different pinouts so it still seems very fishy, from a Japanese manufacturer's POV - all the more since RS is a trademark of Radiospares (UK) and I have seen it on their house branded components over the years. Perhaps it was serviced many years ago? I suggest sorting your chosen transistor substitutes according to specified grading and fitting them as best the available Hfes allow. It can't be huge issue if using Fairchild product anyway.
BTW. I have a bit of fun with a Peak DCA 75 myself - lots of DC performance graphs to see on your PC if you need them, but I don't often need that much data.
With electrolytic capacitors, because they do all dry out and corrode to useless scrap over time, the shotgun approach is indeed what you do. Just replace them all after listing from the parts list and buying the values needed. After so many years, the available values/ratings may be different now but there is no problem using higher voltage ratings or up to 2x the capacitance values in most low-level audio circuits. The important issue is to do it - there are too many to consider individually. You won't need a capacitance or ESR meter to prove anything when they are pulled - simply bin and replace.
I don't know what to say about the BC109s - they are plain vanilla, ungraded types and at odds with the original specs. - Id say it was an unplanned substitution whoever did it. BC109s are European (prolectron) types with different pinouts so it still seems very fishy, from a Japanese manufacturer's POV - all the more since RS is a trademark of Radiospares (UK) and I have seen it on their house branded components over the years. Perhaps it was serviced many years ago? I suggest sorting your chosen transistor substitutes according to specified grading and fitting them as best the available Hfes allow. It can't be huge issue if using Fairchild product anyway.
BTW. I have a bit of fun with a Peak DCA 75 myself - lots of DC performance graphs to see on your PC if you need them, but I don't often need that much data.
Thanks Ian, for your advice about the caps. I wasn't aware of their relatively short life expectancy - but it all makes sense!
Re: the RSBC109's at TR9 - yes, their contacts were indeed twisted to get their E.C.B. leads in the right holes - 'whoever-it-was' had put insulation on one lead where they crossed.
Caps. I frequently change these in Graphics Cards in PC's, and, Yes, I was just looking at the price of the PEAK ESR70 but your timely comment has just saved me £84. Thanks! I shall make a caps shopping list and replace all on the Pre-Amp PCB. I haven't yet looked at the System Control Board (TH-1015S) PCB with a view to changing any components but there are 32 transistors including a number of 2SC945's - yet I don't think they are suspect transistors (are they?) There are 57 diodes of various types - silicon, germanium and Zeners (again - not suspect - are they?) and there are 57 or so caps.
So should I concentrate on changing ALL the caps on both boards? but just the recording chain 2SC458's to KSC1845's?
This is quite a mammoth undertaking - but I'm retired and have the time and shall not do anything until I'm happy that I'm doing the 'right thing'.
Now I must do some 'Cello practice for my Piano trio on Wednesday . . . .
Michael (UK)
Change the electrolytic caps only at first. The other components, short of random failures, could last lifetimes before corrosion gets them or accidents blow them. By all means, start at the input preamplifiers and work your way through to the playback amplifiers, board by board, replacing caps as time and money are available. Control circuits may not affect the sound as long as they operate correctly but will still need routine replacements at some time soon. Replace the suspect transistors too, if you wish but if they are functional, that's not essential.
Having said that, the sheer number of semis in there is always a threat from a statistical POV. It may only take one switching or signal diode failure to kill the sound but you can trace problems in AC coupled circuits like these fairly simply with a signal generator or any low level, constant audio tone or sound level. Simply proceed from the output end back to input, node by node, until the audio at the ouput cuts out and you have found the fault area, be it a switch of electrical or mechanical nature or a dead transistor, capacitor etc.
There are "how-tos" on the web to help you identify specific component faults without spending multiple sums for various special instruments. A good but often inexpensive DMM is now an accurate and flexible device for many component tests, just as it comes. If you don't already have one, this is indispensable for just about all electronic measurements.
Having said that, the sheer number of semis in there is always a threat from a statistical POV. It may only take one switching or signal diode failure to kill the sound but you can trace problems in AC coupled circuits like these fairly simply with a signal generator or any low level, constant audio tone or sound level. Simply proceed from the output end back to input, node by node, until the audio at the ouput cuts out and you have found the fault area, be it a switch of electrical or mechanical nature or a dead transistor, capacitor etc.
There are "how-tos" on the web to help you identify specific component faults without spending multiple sums for various special instruments. A good but often inexpensive DMM is now an accurate and flexible device for many component tests, just as it comes. If you don't already have one, this is indispensable for just about all electronic measurements.
Last edited:
It transpires there are only 6 electrolytic caps on the Sys.Con.PCB and 3 on the Pre.Amp PCB. All the rest are a mix of Metal Oxide Film, Metallized Film, Tantalum, NP (whatever that is) and Styrol capacitors. So my order is quite short - just 9 Elec. Caps.
I got an email from DIY Audio about Ian's latest post but as yet nothing posted. This is an asnwer to that email.
Michael (UK)
Thanks for your input, Jerlywoo. Please would you advise: Are both the NP and the Tantalum non-polar? i.e. I don't need to bother which way the replacements go in ?
And:
As for replacements should these capacitors be as per the original - that is to say should a replacement Tantalum also be a Tantalum, and the NP/C as NP/C? Are these still available?
By this you can tell I'm more at home tuning a piano for a recital . . . !!
Michael (UK)
So for my 2 Tantalum caps I should order standard Electrolytic caps (with the same 47mF and 100mF values and observe the polarity? The voltages are 6.3WV and 3.15WV. How tolerant are Electrolytics to voltage? I believe I can use higher voltage OK (but not lower!) Michael (UK)
There's something of a consensus here that 16V is a reasonable min. voltage rating to use generally because it tends to lead to better reliability and you can still read the labels! Parts have shrunk to the point where the size difference is enough to use a blanket minimum rating of 16 or 25V and this simplifies purchasing. You also need fewer spares.
We haven'discussed grades and this a big deal for some guys but it won't hurt here to use industry standard, low ESR grades like Panasonic FR, FC, FM or equivalents in other respected brands like Rubycon, Nichicon, Nippon Chemicon, Elna etc. Some distributors stock brands you don't find elsewhere and obviously, you take your chances with them.
We haven'discussed grades and this a big deal for some guys but it won't hurt here to use industry standard, low ESR grades like Panasonic FR, FC, FM or equivalents in other respected brands like Rubycon, Nichicon, Nippon Chemicon, Elna etc. Some distributors stock brands you don't find elsewhere and obviously, you take your chances with them.
I think I should have started a New Thread for this discussion of 'Capacitor replacements'. It's a totally different subject from the 'Transistor substitutes' I started. My apologies! But thank you for this concensus! I am now creating a purchase order for caps based on suggestions from this site. Michael (UK)
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.