Yeah, I have BD139's here and they're a no-go for this.
Thx for bringing it up, the 6144 is an interesting device and could be a great choice. I'll get some next time I order.
What D882 has going for it is for the beta to hold up well at a fraction of a volt Vce. It does start to drop below 5 volts on the TIP. At low supply voltages, this matters.
Modern TIPs can be very consistent unit to unit, same as any other type made these days. But don’t expect the same consistency lot to lot. You may not even get exactly the same die, if the same device “type” is processed at multiple fabs.
Today I put a matched pair of the TTC004B into one channel. After 30 minutes, temps on those transistors (via infrared thermometer) are running 51ºC, which is 1-2ºC hotter than the 41C's on the other side, and their heatsinks about the same amount lower. So heat tranfer is less good but good enough.
I measured THD+N at 1kHz and 1VRMS into loads that are representative of the headphones I use the most: 220Ω and 44Ω. At 220Ω, THD dropped 1dB from -98 to -99 dBV. Into 44Ω, it dropped from -88 to -91. So the extra current gain makes a difference when it's needed.
Listening using mono source material and one channel driving both headphones, the side with the TTC004B sounds smoother and more effortless. By comparison, there is a pinched or thin quality in the 41C side.
I consider this a success and will upgrade the other channel tomorrow.
Thanks to everyone for your input.
I measured THD+N at 1kHz and 1VRMS into loads that are representative of the headphones I use the most: 220Ω and 44Ω. At 220Ω, THD dropped 1dB from -98 to -99 dBV. Into 44Ω, it dropped from -88 to -91. So the extra current gain makes a difference when it's needed.
Listening using mono source material and one channel driving both headphones, the side with the TTC004B sounds smoother and more effortless. By comparison, there is a pinched or thin quality in the 41C side.
I consider this a success and will upgrade the other channel tomorrow.
Thanks to everyone for your input.
I got the 2SC6144's in today. They too are fully encapsulated. Not what I expected.
The datasheet says 2SC6144 is a TO-220F package (sometimes denoted as TO-220FP) meaning the package is fully insulated and don't need any additional insulation and so can be directly mounted onto the heatsink, but due to the insulation being same material as the rest of the black plastic encapsulating the die the thermal resistance is very poor, for example STP12PF06 and STF12PF06 (datasheet attached) are both the same MOSFET but the power rating is 225W for TO-220 and 60W TO-220FP due to Rthj-case difference 2.5 VS 5.35 °C/W.
The encapsulated transistors are likely 1 to 2 watt packages.
Be better to use a 7 to 15 watt packages that have thermal pads.
KSA1381 / KSC3503
T0-126 = 7 Watt ECB - Rbjc = 17.8 C/W
fT = 150 Mhz
Hfe = E suffix 100 to 200 - F suffix 160 to 320
------------------
MJE243 / MJE253
T0-225 =15 Watt ECB - Rbjc = 8.3 C/W
fT = 40 Mhz
Hfe = 40 - 180
Far as High frequency performance
I think 120 ma is rather high.
You likely dont need much more than 30 to 40 ma
to get all your gonna get at high frequency.
anything higher dont do much and just putting heat into the devices
Be better to use a 7 to 15 watt packages that have thermal pads.
KSA1381 / KSC3503
T0-126 = 7 Watt ECB - Rbjc = 17.8 C/W
fT = 150 Mhz
Hfe = E suffix 100 to 200 - F suffix 160 to 320
------------------
MJE243 / MJE253
T0-225 =15 Watt ECB - Rbjc = 8.3 C/W
fT = 40 Mhz
Hfe = 40 - 180
Far as High frequency performance
I think 120 ma is rather high.
You likely dont need much more than 30 to 40 ma
to get all your gonna get at high frequency.
anything higher dont do much and just putting heat into the devices
Last edited:
I'm really sorry for that. I didn't have the package in mind when I recommended the transistor. I haven't even had a close look at mine, kept the bags sealed.
Since you only dissipate 1.4W, poor Rthjc is not a big issue in your application.
In case the Rthjc is 5 K/W and you have a small radiator with say 20 K/W, the silicon is at ~50° then.
Higher temperature also increases hfe, which is what you are looking for.
When I bought mine, I had ~1A bias in mind and 12-15V supply so I need to reconsider the application.
Does anybody know what is inside the fully encapsulated packages? I guess that the silicon is still mounted to a slab of metal that is then encapsulated? If so, could it be an option to just sand away the plastic at the rear in order to expose the metal slab?
It’s the same lead frame as a regular TO-220. I suppose it’s possible to sand away the encapsulant, but flatness and surface roughness may be a problem and you’d be back to poor heat transfer. Sand then polish, maybe. You may be just as well off buying a surface mount version and soldering a copper heat spreader (with a mounting hole) to it. Either way a lot of work, probably best off just buying something else and hanging on to these to use where they are suitable. In the case of 1.4 watts, they would be just fine.
Unfortunately, they don’t give you very many options for non-encapsulated TO-220 “driver” transistors anymore. And you don’t always want (or need) the 8-amp MJE1503x, and you may very well want something faster.
Unfortunately, they don’t give you very many options for non-encapsulated TO-220 “driver” transistors anymore. And you don’t always want (or need) the 8-amp MJE1503x, and you may very well want something faster.
No problem and not your fault, Lee. I should have noticed but was concentrating on other details.
Agreed, it would work fine for this project. The TTC004B is also working fine (and sounding fantastic, BTW) in it, so I'll leave it be. I actually got the 6144 for a different Class A project, running 12V @ 350-400mA bias with a very generous heatsink.
I have KSC3503 here, it is an ok driver but has significant beta droop above 50mA, so no good here. And good luck finding an F-suffix part! When I bought 2 years ago, I only found D parts in stock. The MJE243 might be ok but isn't an improvement over the TIP41C in Hfe.
Hello,
BTW, has anyone used the TTA/TTC004 in Ltspice? For me, the Toshiba-Parameters look quite strange with very low early-voltage? My measurements are around 800@10mA (cannot messure exaxctly at DC-conditions, but its „high“).
Regards, J-C
BTW, has anyone used the TTA/TTC004 in Ltspice? For me, the Toshiba-Parameters look quite strange with very low early-voltage? My measurements are around 800@10mA (cannot messure exaxctly at DC-conditions, but its „high“).
Regards, J-C
I generally use Bergquist Sil-Pads, without grease, on TO-220s. They correspond to plotted lines (5) and (6) in the figure below, lifted from ON Semi appnote AN-1040/D.
_
Attachments
I used TTC004B/TTA004B as driver in 120W/4ohm amplifier to drive 2xTTC0002/TTA0002 outputs. It is fine, runs since approx. one year, with no problem. The drivers are mounted directly to the heatsink, without additional insulator. Bias is 25mA for the TTC004B/TTA004B pair.
That's the very next chart in the On Semi app note, whose identifier I mentioned in post #33
The answer is - LARGE impact. The Sil-Pads would make a lot of sense in a production environment, though.
Since Junction to Case is 12.5K/W here and no insulation is needed, I do not expect any significant difference here if the mounting of the device is done proberly. But as always, I would keep a factor of ~5 for continous dissipation to the max values in the data sheet.
Means, for more than 2W (on heatsink), I would choose a bigger device. And from ~ 300mW on I would recommend heatsinking.
Regards J-C
Means, for more than 2W (on heatsink), I would choose a bigger device. And from ~ 300mW on I would recommend heatsinking.
Regards J-C