Radio sound quality - some measurements

[GeoffK]

I do not spend all my time looking at the forum so have only just read this, as it happens I didn't think the parallel components left in place would have a large influence on the overall frequency response or its dB points. As Radio Wrangle points out, it is a fairly complex circuit to determine exactly what is going on, the main determinate of the high frequency cut being the components in the tone control, that even after removing the loudness components still gives adequate range. Clearly posts of this nature attract a large audience but few contributors. I am sorry if you though I was talking bull**** as your quote implied. I should also point out that with the components removed the volume control at its minimum and lower setting is still quite controllable, it does not go from zero volume to very loud with a small movement of the volume control. I carried out this mod to my radio several years ago, before this forum was started and have been very happy with the performance of the radio since then. It is helpful that mhennessey found an alternative easier way to achieve a modification to the loudness.

[Radio Wrangler]

I've just realised there was an ambiguity in my post B word = Bose, just to clarify.

David

[mhennessy]

Originally Posted by Radio Wrangler

That is a complicated circuit with several time constants close enough together to interact, so a normal manual calculation is either going to be a very involved full analysis, or the usual bode-plot approximations are going to give errors about as big as the amounts of shaping in the midband.

You're right; it is complex. Certainly, I wouldn't want to produce a complete analysis of the whole circuit by hand, but here, we just had to approximate the difference between 2 components fitted and not fitted, and it only took 30 seconds to come up with a ballpark figure for the change at 10kHz...

My back-of-the-envelope approximation was crude: just using a bit of Ohm's law and making a few assumptions allowed me to simplify the circuit and get to a point where I could see a worst-case change of about 2dB. Refining the simplifications - or running a simulation or real test - altered that figure by only a small amount. Often, the rough answer is close enough, providing you were lucky with your assumptions...

But it gets really interesting when you consider the turnover frequencies of the capacitors. It doesn't take long to realise that they are turning over in the sub-1kHz region, just as the plots show. So it's not about the HF after all

I was never a fan of Grundig, they seemed to do too much messing around. I suppose a bit like the Bose radios of their day, although without that level of marketing bull. Oh crikey! That's torn it... I've mentioned the B-word!

Generally, Grundig did a good job with the equalisation on the smaller sets, but larger sets like the Yacht Boy seem to suffer from over-blown bass (presumably that was their "house style" - perhaps to help sell the large radios over the smaller ones?).

But "fixing" the N210 was easy - and worthwhile, given how ubiquitous they are. I've lost count of the number of discussions about Yacht Boy Bass we've had on this forum alone, so it's good to have some objective measurements on record, along with some options for those who wish to experiment

As for Bose: Buy Other Sound Equipment

[stevehertz]

Personally I love the 'Grundig sound'. But in any case, they have tone controls there for us all to use, if there's too much bass, just turn it down! But they don't normally have a strident, peaky upper mid range as do a lot of portables. Also, having been a musician (in bands, solo) all my life, I do like a 'big sound', one that sounds more like a hifi (as much as one can expect) than a typical tinny portable. Not that Hackers and the like are tinny.

[mhennessy]

But the trouble with the Yacht Boy specifically is the bass control is a simple switch which cuts the bass far too aggressively.

All the other sets of that size that I've seen have a variable bass control, luckily.

My mod involves shorting out one capacitor, which leaves it with a maximally flat response, but a pot could be used to get variable bass lift. I'm already experimenting with that...

To be honest, the sharp roll-off at 10kHz is more of a problem. I don't think the speaker has much left to give, based on a test with my Clarke Teknik EQ, so it's looking like a different speaker or a tweeter - neither option is easy or appealing right now. Still, I won't care about 10kHz in a few years time

[stevehertz]

Yeah, when I play my test CD, the old 10KHz test tone doesn't seem much more than a gummed up dog whistle these days. But hey, I know my mid range..

[Philips210]

Hi.

Something that has occured to me is how will the frequency response be affected with wear to the loudspeaker? Does it remain fairly consistent over the years or is there a subtle change due to wear in the cone and cone suspension. Does the composition of the paper cone degrade thus changing the frequency response. Humidity may also play a part in altering the acoustic properties of the speaker under test. No doubt the acoustic transparity of any lining cloths or speaker cloths may affect the overall response.

Regards
Symon.

[Radiocruncher]

So, the R606 measures (and sounds) better than the R606MB. Who knew

This makes for very interesting reading Mark and thanks for taking the time to carry out the tests and share the findings.

I was interested to see the difference between the 2 R606 versions but the graph seems to be missing.

Regards

Graham

[GeoffK]

Originally Posted by Philips210

Something that has occured to me is how will the frequency response be affected with wear to the loudspeaker? Does it remain fairly consistent over the years or is there a subtle change due to wear in the cone and cone suspension. Does the composition of the paper cone degrade thus changing the frequency response. Humidity may also play a part in altering the acoustic properties of the speaker under test. No doubt the acoustic transparity of any lining cloths or speaker cloths may affect the overall response.

High humidity is probably the enemy of speakers, apart from too many Watts damaging the coil, not least as rust can set in to the metal parts. Some speaker can suffer from the cone sagging, but paper seems very durable if used in good conditions. High volume settings can loosen up the suspension and cone surround that could alter the response of the speaker. But if looked after they will last indefinitely, even the magnets loose very little of their magnetism.

[mhennessy]

Originally Posted by Radiocruncher

I was interested to see the difference between the 2 R606 versions but the graph seems to be missing.

Oops! Sorry about that...

[Radiocruncher]

Thanks Mark. I will have to dig them both back out for a listen.

Regards

Graham

[mhennessy]

Over on this thread, I've posted results from the Roberts RM30/33, along with some suggested mods to improve the performance. While I'm at it, I decided to measure very similar RCM1, but I've put the results here to keep that thread on-topic.

Electrically, the RCM1 is an RM30 with an additional clock module with a nice deep VFD with large digits. The case is obviously wider to accommodate that, but also it's slightly taller. This allows for a larger loudspeaker - this one is 4" round. Apart from the size difference, it appears to be the same as the smaller elliptical units.

This one is more extended at the LF end, and seems to hold up better at HF too. The graph shows this (in blue) against the original RM30 (black). Remember - the electronics is identical between these two sets, so the differences are down to the loudspeaker+cabinet.

This set obviously sounds a bit better than the smaller RM30/33. While I haven't tried it, I suspect the mods would work well, but using the same values would result in too much bass and treble, I suspect. So less EQ will be needed, meaning there will be less loss of volume overall. If I'm deluged with requests, I'll try it at some point in the future.

Also, it's possible that this model also exists in TDA5700 form, and if so, it's likely to be a bit brighter, as we found with the RM30/33 models.

It's also worth noting that loudspeaker upgrades might be possible (though the mains transformer is almost touching the drive unit, so that'll limit the options).

Cheers,

Mark

[mhennessy]

And just to finish this run of table models from Roberts, here's the RM50.

I admit that this one is better than I'd remembered. It was one of the first radios I bought, but as I was anxious to make a start on my new Hackers, this one was put aside for a rainy day. It should restore nicely, if I ever find the time

This uses a 6 by 4" speaker, which looks very similar to the other types seen in the RM30/33 and RCM1. The power amplifier is fairly meaty, using the AD161/162 pair to deliver some 5 watts. The bass and treble controls use the active Baxandall arrangement - like the R707 - but unlike the R707 there's no loudness tapping.

There are two plots - the first (black) shows the response with the tone controls in their centre position, and the blue plot is taken after they have been adjusted for the best sound (bass at 4 o'clock, treble at 3).

In common with the other table models, the treble is very limited. But the bass is better, peaking at the resonant frequency of ~100Hz before falling away rapidly. Careful with the bass control, as like the R707, it's easy to make this set go boomy! But unfortunately, the treble control brings up the midrange hump at 3kHz, which makes this set somewhat painful to listen to. Ideally, it needs a lift starting at 6 or 7kHz to bring up the high treble region. Looking at the schematic, I see that they've split the emitter resistor of the preamp into two (R35 and R36) for no apparent reason, and I'm wondering about adding a capacitor in parallel with R36 to get some upper HF lift. That'll have to wait for another day...

This set is a bit more spacious than the others, so a drive unit upgrade is quite feasible.