-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Topping PA5 II Stereo Amplifier Review
- Thread starter amirm
- Start date
I couldn't find a spec but I'll bet it's somewhere around 10k to 22k ohms. I would plan on it being 10k ohms, as that seems pretty standard these days. What will you be driving it with? The output impedance of your sources should be no higher than 1k ohms. Almost all audio sources have low output impedance these days, but there are some analog preamps that have higher output impedance, especially vintage vacuum tube phono preamps.
Those tiny Bose cubes made tower arrays sound small and were the ultimate 90's status symbol...
daniboun
Major Contributor
If same as PA5 : 10K
Last edited:
Thanks Rongon and danibound for the info. Maybe it's not safe for my case: I do have a lovely Sansui Tu-517 tuner. It's only output is variable, and I guess its impedance must be quite high. I am not certain, but I do own an Accuphase T-101 which, according to the manual, has around 2k in its variable output (and only around 200 in the fixed output).
Well your Sansui is 600 ohm output impedance, so that is fine. And a higher impedance on a variable output is less of a problem as you'll adjust the ouput level to compensate the voltage drop
Last edited:
Oh, that's good news then. Thanks!
The 'rule of thumb' concerning matching components' impedances is that ideally the load (target) impedance should be 5X to 10X the source impedance.
The larger the ratio of Zout:Zin the better the signal voltage transfer, least THD, and widest bandwidth. But 1:5 (2k:10k) should be okay.
The larger the ratio of Zout:Zin the better the signal voltage transfer, least THD, and widest bandwidth. But 1:5 (2k:10k) should be okay.
As long as both load and source are purely resistive (they should be), there should be no impact on THD or frequency response. Only signal level should be impacted by a relatively high output to input ratio.
The thing is, the signal source will usually be AC coupled (have an output capacitor to block DC offset), which will form a high pass filter with the load impedance. This will be especially likely with a source having a vacuum tube output stage, as these will usually have low value output DC blocking capacitors with a value of 2.2uF or possibly less. (They need to be rated for at least 250V DC working voltage, and they're usually film capacitors as opposed to electrolytic types.) It is true that a DC-coupled output stage with near-zero DC offset shouldn't exhibit any reduction in low frequency response relative to mid-frequency. But how many vintage analog sources have a DC-coupled output stage?
Also, many line level amplifier stages will exhibit increased THD driving a heavy load as compared to how they'll perform lightly loaded. This is likely to be worse in a signal source that has high Zout, such as a vintage analog signal source like a tuner or phono preamp, and especially if those have a vacuum tube or discrete transistor output stage (as opposed to a modern op-amp). It is true that not all signal source output stages will exhibit markedly higher THD when heavily loaded. But many will.
Also, many line level amplifier stages will exhibit increased THD driving a heavy load as compared to how they'll perform lightly loaded. This is likely to be worse in a signal source that has high Zout, such as a vintage analog signal source like a tuner or phono preamp, and especially if those have a vacuum tube or discrete transistor output stage (as opposed to a modern op-amp). It is true that not all signal source output stages will exhibit markedly higher THD when heavily loaded. But many will.
Last edited:
Every day is a school day. You are correct, I'd not considered the impact of any DC blocking capacitor.
Lets put to one side tube based designs. Niether of the tuners mentioned are tube based, and if someone is selecting tube based pre-amplification, we can probably assume they are not that worried about frequency response and distortion compliance.
Then FR flatness will depend on the ratio of total resistance (Output + input) to reactance of the blocking cap. In this case either of the resistors increasing in value will tend to flatten the FR. So Ouput R = Input R would have a flatter FR (Although at a lower level) than Ouput R< 1/10 Input R.
Similarly a high output R will actually reduce the load on the line level amp. I'd also hope that 20K would not be considered a significant load even for a vintage line stage. Though I have little knowledge of what such a vintage output may look like, so that hope might be misplaced.
1. Most AM/FM tuners are going to be vintage, in the sense that they're likely to be 1970s or '80s designs, and so will have output capacitors and possibly high output impedance. Back in the early 1980s, audiophile designs took great pains to never use op-amps in the audio stages, even though the NE5532 was available and was really quite good already. TL072 was available back in the 1970s, but again, many hi-fi mfgs would use discrete designs, which often had higher Zout than we're used to seeing today. Note that none of this is vacuum tube gear.
2. The relationship of Zsource (Zout of source) to Zload (Zin of load) is not going to have influence on the frequency response (FR) by itself. The relationship of the output blocking cap to the Zload is what will form the CR high-pass filter (HPF). That will dictate the F3 of the HPF. It doesn't matter what the Zsource is in that equation.
A really high Zsource can interact with cable capacitance and the load's input stage capacitance to form a low pass filter (LPF). That's because the cable capacitance appears in parallel with the resistance and capacitance of the load (power amp input stage). For example, if we have a source with 2k ohm Zout, feeding a load (power amp input) with 500pF input C ("Miller" capacitance), and a long cable with let's say 2.5nF (2500pF) capacitance, that's like an RC filter (HPF) having a 2k ohm series resistor and a 3000pF parallel capacitor. The F3 of that LPF would be uncomfortably low at 26.5kHz -- almost down to the audible range. As you can see, all this has nothing to do with the Zload or input resistance of the load (power amp).
3. A high output Z (Zsource) working into a low value of Zload will more heavily load down the source device's output stage. For closest to ideal performance, we want to more lightly load the source device's output stage -- reduce the load -- not load it more heavily. That's all terminology stuff, but I figured I should clarify how I'm using these terms.
Rules of thumb:
1. You want to load the source's output stage as lightly as possible (which means make the ratio Zsource:Zload as large as possible, i.e., 1:10, and not 1:2 or 1:1).
2. You want to keep cable runs as short as possible, to avoid introducing more load capacitance (thus loading down the source at high frequencies).
3. You want to keep the source's output stage impedance (Zsource) as low as possible, so it can maintain good bass response into all possible loads (including bridged Class D amps with Zload of <5k ohms, etc.).
4. You want to keep the load's input stage impedance (Zload) as high as possible, so source components can work well at all frequencies (including the lowest freqs) even if they have weedy high impedance/low current output stages.
I haven't gotten into the signal source's ability to sink current into a load and the possibility of slew rate limiting because a) I doubt it's an issue in this case, and b) it's a bit complicated and would take a lot of explaining. Suffice to say that another rule of thumb is that a signal source should be able to source enough current into the load so that it doesn't slew limit at high frequencies into a high load capacitance (such as would be encountered if you use long runs of cheap 'n nasty interconnect cables). However, that's an extreme case, so I won't belabor that point. Most contemporary op-amps (OPA2134, LM4562, NE5532, etc.) can supply 5mA or so of signal current (unbalanced), which is plenty for any reasonable load capacitance.
I keep using Z for impedance rather than R for resistance because I'm talking about AC audio signal voltages here, not DC voltages. Since audio signal is an AC phenomenon, and impedance can be defined as 'frequency-dependent resistance' or a resistance that's likely to vary with frequency, I think using Z is more appropriate (less confusing) than using R in this particular discussion.
Sorry about the length of this post. I guess I got carried away...
Last edited:
Just for fun, here is the output stage circuit of the Sansui I found in a service manual.
That is fun!
The output stage uses a 2SC945 with the output DC coupled from its collector, which would be its high impedance output (as opposed to its emitter, as in an emitter follower, which would have very low output impedance).
However TR07 looks like the 'real' output device, as it's an emitter follower with a 10uF capacitor (C21) on its output, driving a 10k ohm load resistor (R31). I think TR09 is an active load for the emitter of TR07, to look like a very high impedance load, possibly isolating it from low impedance loads downstream.
I have to admit that I don't have the chops to just look at that circuit and see exactly how its output stage is working, especially with so few voltages shown. But I do have LTspice at home, which can do the calculations for me. I could simulate it and test it virtually. That would tell me if it would load down appreciably into a heavy load like 10k or 5k ohms.
Like I said.... Fun.
PS - It looks like LTspice has a model for 2SC945, but it does not have a model for 2SC1222. I'm hoping 2N2222 would work. We shall see...
PPS - It looks like BC550 is a near equivalent of 2SC1222. Perhaps close enough for in-the-ballpark figurin'.
The output stage uses a 2SC945 with the output DC coupled from its collector, which would be its high impedance output (as opposed to its emitter, as in an emitter follower, which would have very low output impedance).
However TR07 looks like the 'real' output device, as it's an emitter follower with a 10uF capacitor (C21) on its output, driving a 10k ohm load resistor (R31). I think TR09 is an active load for the emitter of TR07, to look like a very high impedance load, possibly isolating it from low impedance loads downstream.
I have to admit that I don't have the chops to just look at that circuit and see exactly how its output stage is working, especially with so few voltages shown. But I do have LTspice at home, which can do the calculations for me. I could simulate it and test it virtually. That would tell me if it would load down appreciably into a heavy load like 10k or 5k ohms.
Like I said.... Fun.
PS - It looks like LTspice has a model for 2SC945, but it does not have a model for 2SC1222. I'm hoping 2N2222 would work. We shall see...
PPS - It looks like BC550 is a near equivalent of 2SC1222. Perhaps close enough for in-the-ballpark figurin'.
Last edited:
you'd have to have a bloody long cable to get that at a typical 50 to 60pF / m (yes you made me look it up)
I'd put it to you that cable capacitance is not going to be an issue for most people in real systems - even with a 2K output impedance.
I'd also be surprised to find an input stage (at least in a modern design) presenting a significant capacitance directly on the input stage without an intervening buffer.
I've measured some cheap 'n nasty RCA cables that measured 100pF a *foot*. Remember that the plugs also add capacitance.
Also, the input wiring of an amp will add a few picofarads. It shouldn't add any more than that, but...
But yes, I was being a bit over the top dramatic. Still, 12 feet of 100pF/ft would be 1200pF (1.2nF). It's not *likely* but it is *possible*.
Class D amps have an input buffer, but most Class AB amps do not. It's fashionable in some circles to use JFET input stages which do have high input capacitance. But again, I was being overly dramatic to make the point. Again, a situation that's not likely, but is possible.
PS - When designing something, I think it best that it be overbuilt. You want your widget to sail through any torture test that might be possible in the real world. If you design a widget to be just adequate for intended use, somebody will use it outside its comfort zone and wail about it online somewhere...
Also, the input wiring of an amp will add a few picofarads. It shouldn't add any more than that, but...
But yes, I was being a bit over the top dramatic. Still, 12 feet of 100pF/ft would be 1200pF (1.2nF). It's not *likely* but it is *possible*.
Class D amps have an input buffer, but most Class AB amps do not. It's fashionable in some circles to use JFET input stages which do have high input capacitance. But again, I was being overly dramatic to make the point. Again, a situation that's not likely, but is possible.
PS - When designing something, I think it best that it be overbuilt. You want your widget to sail through any torture test that might be possible in the real world. If you design a widget to be just adequate for intended use, somebody will use it outside its comfort zone and wail about it online somewhere...
Nor do I, but I have the full circuit in the service manual. I think TR 5 6 9 and 10 are part of a mute circuit (called pop protection). Basically switch them on and nothing comes out.
Circuit is confusingly drawn. They make it look like a push pull output stage, but it is two independent outputs drawn as an upside down mirror image.
If. you are interested, you can get the service manual here - you might need to create a free account.
Sansui TU-517 HiFi Stereo AM/FM Tuner Manual | HiFi Engine
I received PA5 II Plus two day ago. After short listen to make sure that everything is fine, typically for me, I opened it See some pictures below. Not sure if you'll see anything unseen in previous posts, but here it is some info and my personal opinion on this little amp.
My overall impression is of very good build quality. SMT solder joints are nice, although skimpy on solder, which is the usual situation nowadays. Hand soldering is also clean, with a little remark that the solder didn't flow on both sides of the board. Example of that can be seen on the output jack wires - the solder on the bottom side is great, but no solder went through the via on the top side. For someone coming from aviation and aerospace that should be a poor solder joint and a "fail", but I guess for audio is not that crucial. However, these are places with potentially high current, so a little better solder flow wouldn't have hurt. The PCB is gold immersion which is impressive for the price!
Components are of good quality - same stuff that you'll find on Digikey etc. Nothing looking outright cheap. Electrolytic caps are rated for 105C from what I saw. This is important considering the potential heating in the enclosure. Reliability should be good in that respect. There is one 5532 and one 1612 opamp per channel. I suspect one of them works as buffer for the volume pot, creating "active" volume control. Then the potted cans contain the preamp section, composite amp etc.
Heatsink is small in my opinion, but probably okay for very efficient class D operation and no big house party style abuse. I couldn't find any data on the TPA chips stating efficiency, so I'm not sure if the heatsink is adequate for full continuous power. Most likely it's not! I think the idea that this thing can run 2 x 120W continuously is ridiculous. As Amir mentioned, his power measurements are several cycles sweeps. I would say this qualifies rather as instantaneous power than continuous. In addition the PLUS power supply is 38V/4A, so no way to achieve 2x120W/4R for more than couple of hundreds milliseconds. Only 2200uF supply filtering cap doesn't help either. Having said that, I don't really care. In my case I'm fine with much less. I've never had the need, and never owned, an amp of more than 2x50W. My listening impressions are that this little amp can do 2x20W realistically (maybe), but I may be wrong.
I read some posts mentioning that the enclosure works like a heatsink. This is wrong. The heatsink doesn't connect to the case in any way. So, if the enclosure is getting hot, this is heat convection, not conduction. My PA5 II doesn't get hot in stand by - maybe 35C max in normal 25-ish degree room. It warms up in operation though. I think the idea to change the rubber feet with taller ones is good... The way I listen does not require that.
My subjective listening impressions are: The amp is absolutely quiet and sounds very, very clean. Because it's so clean I find it a little matter of factly, and because it doesn't feel very powerful some people may find it "clinical" or "flat". I find it accurate. I need to play a little with repositioning my speakers. My main class-B MOSFET output amp sounds more expansive left and right, and similarly clean. But is also much bigger. This small box is amazing and I think that it will be worthy replacement for any amp that is of comparable power. If you like "lush", or "rich", or "juicy bass" you won't find that in PA5 II. It will give you the information as is. For lush, rich, or juicy play with your room or change your speakers. My 2 cents...
See some pictures below. Not sure if you'll see anything unseen in previous posts, but here it is some info and my personal opinion on this little amp.My overall impression is of very good build quality. SMT solder joints are nice, although skimpy on solder, which is the usual situation nowadays. Hand soldering is also clean, with a little remark that the solder didn't flow on both sides of the board. Example of that can be seen on the output jack wires - the solder on the bottom side is great, but no solder went through the via on the top side. For someone coming from aviation and aerospace that should be a poor solder joint and a "fail", but I guess for audio is not that crucial. However, these are places with potentially high current, so a little better solder flow wouldn't have hurt. The PCB is gold immersion which is impressive for the price!
Components are of good quality - same stuff that you'll find on Digikey etc. Nothing looking outright cheap. Electrolytic caps are rated for 105C from what I saw. This is important considering the potential heating in the enclosure. Reliability should be good in that respect. There is one 5532 and one 1612 opamp per channel. I suspect one of them works as buffer for the volume pot, creating "active" volume control. Then the potted cans contain the preamp section, composite amp etc.
Heatsink is small in my opinion, but probably okay for very efficient class D operation and no big house party style abuse. I couldn't find any data on the TPA chips stating efficiency, so I'm not sure if the heatsink is adequate for full continuous power. Most likely it's not! I think the idea that this thing can run 2 x 120W continuously is ridiculous. As Amir mentioned, his power measurements are several cycles sweeps. I would say this qualifies rather as instantaneous power than continuous. In addition the PLUS power supply is 38V/4A, so no way to achieve 2x120W/4R for more than couple of hundreds milliseconds. Only 2200uF supply filtering cap doesn't help either. Having said that, I don't really care. In my case I'm fine with much less. I've never had the need, and never owned, an amp of more than 2x50W. My listening impressions are that this little amp can do 2x20W realistically (maybe), but I may be wrong.
I read some posts mentioning that the enclosure works like a heatsink. This is wrong. The heatsink doesn't connect to the case in any way. So, if the enclosure is getting hot, this is heat convection, not conduction. My PA5 II doesn't get hot in stand by - maybe 35C max in normal 25-ish degree room. It warms up in operation though. I think the idea to change the rubber feet with taller ones is good... The way I listen does not require that.
My subjective listening impressions are: The amp is absolutely quiet and sounds very, very clean. Because it's so clean I find it a little matter of factly, and because it doesn't feel very powerful some people may find it "clinical" or "flat". I find it accurate. I need to play a little with repositioning my speakers. My main class-B MOSFET output amp sounds more expansive left and right, and similarly clean. But is also much bigger. This small box is amazing and I think that it will be worthy replacement for any amp that is of comparable power. If you like "lush", or "rich", or "juicy bass" you won't find that in PA5 II. It will give you the information as is. For lush, rich, or juicy play with your room or change your speakers. My 2 cents...
Received my Pa5 II plus version on Monday. So far made a little comparison with LA90 and can't hear any difference. No scientific test of any kind, can't dB match levels or make it a blind-test though..
Question: Does it accept 5v from Topping E70 DAC? Or will it clip?
Question: Does it accept 5v from Topping E70 DAC? Or will it clip?
Last edited:
Similar threads
