Quad 306, an underrated masterpiece?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Member
Joined 2010
Paid Member
Hi Chris. Those pics are simply web sourced. Actually, I replaced the original Cornell-Dubiliers in mine not long after I bought it 5 years ago but as I occasionally find, good caps do last a long time. After careful listening and looking at the PSU behaviour on a 'scope before and after replacement, I could not detect much difference. It was a waste of money in that case but routine replacement is usually about peace of mind for the future.

Something else, and a nasty problem that I did find on this model, was a design fault where the transformer is mounted only to the PCB. The board is partly buffered with a foam rubber pad but nothing else. When I air-freighted a couple more of these amps from the UK a year later, the boards had flexed and fractured the driver transistor legs, shearing some completely. Foam deteriorates and maybe there was a pad fitted to the top of the case originally but it proved inadequate for both consignments. I used TIP42 replacements as the original 40872 types are not exactly high Ft/linearity types and I don't like provoking instability without a sensible reason to.

The flimsy, hobby grade Deltron speaker sockets also fail and IMO, banana fittings may be traditional in the UK but a bad idea now, for consumer applications. On the positive side, there are no electrolytics in the amplifier apart from the main PSU caps, a great benefit to consistent, long-life performance and that is worth every cent ;)
 
If only the Quad designers took into consideration that they are designing consumer electronics and not lab equipment. No matter how good they sound, I was automatically put off by their looks. I had for one weekend a Quad 303 and Nakamichi PA-7 in my system many years ago, and I much prefered the sound of the PA-7. How much of that was because of me being subjectively biased by the looks I will never know.
 
Member
Joined 2010
Paid Member
It's interesting how looks appeal or not to different age groups, social background etc. I grew up when satin finish, extruded aluminium fascias and lots of knobs everywhere were the latest thing and offered the most impressive looks in Hifi.

By 1987 when the 306 arrived, the fine industrial art looks were well out of date and the colour remains as about the drabbest I've ever seen too. The painted steel of all Quad preamps then, had been passé for front panels since the 1960s. The 306 was also overpriced and scarcely represented here, where few ESLs were to be found to justify buying. I understand very few units were imported.

However at the right price, I find that in a small system the pair does look fabulous in the tradition of fine 20th century industrial design but there needs to be a visual break with your typical black/silver gear. You can't avoid a problem with modern sources like players, receivers, external DACs and their remote controls etc. but it pays not to just slap everything close together and then try to convince yourself (good luck with SWMBO) that it all looks fine.

I think the tilt control is a wonderful idea too and does everything you could need for small tonal balance adjustment with precision. Just take care with the switches. They are of amazing over-designed quality but not robust where it matters and very difficult, if not impossible to repair. If you are lucky, they won't have been used much, meaning they could be fine.
 
Hi all,i am a new member and a new owner of a 306.As i am not very familiar with DIY electronics i would like to ask one or two things.
1.I want to alter the input sensitivity to 1 vol,is the change of the specific resistances the only thing that i have to do ?
2.What type of electrolytic capacitor should i use for the power supply section recap ?
Any other suggestion is more than wellcome.
----------
THEMIS
 
Member
Joined 2010
Paid Member
The electrolytic capacitor proportions (short and fat) mean that there aren't a lot of practical choices for the 4,700uF/50V caps. However, they are at least common snap-in types (ie 10 mm pin spacing) that you can search in any decent supplier's on-line catalogue. As a guide of performance, a 3A ripple current rating would be my choice, as in Nichicon LL type for example. Measure the diameter and length of your caps to ensure any replacements will fit with plenty of clearance.

You don't need to reduce amplifier input sensitivity to accommodate a preamplifier where the output voltage is variable. You do need a preamp with a volume control in any case since a fixed maximum output voltage only applies to players and sources without them. A simple 2-resistor 8 dB attenuator network could be added between preamp source and amplifier, mounted inside the amplifier (and shielded from the power supply) if you wish. Here is a tutorial allowing you to understand and calculate values for yourself: Passive Attenuator Tutorial about Passive Attenuators.
 
Thanx a lot Ian,then one of the following should be OK ( same price 5GBP/p )
1.ELNA 8200/56V LAO High End Audio Electrolytic Capacitor
2.NICHICON GOLD FW 6800/63V Radial electrolytoc capacitor for Audio
I would say the first since 56 volt rating is sufficient.
I have seen far more expensive caps but i dont believe that they will make any difference ...
Both my LINN PRETEK preamp and TRICHORD GENESIS do have output controls but i was thinkig of being on the safe side on matter of distortion(?) .
----------
THEMIS
 
Member
Joined 2010
Paid Member
You can change what you like but consider that your inspiration to use "absolutely best" components in a small class A amplifier won't improve anything because it can only gild the lily. Small signal devices are selected very well in the original design because costs of such devices are minimal and there has been little scope for improvement in small discrete audio semis for decades. Sign up and download the free service data PDF including the schematic here: Quad 306 - Manual - Stereo Power Amplifier - HiFi Engine

The class A amplifier consists of T4,5,6 but note that T5,6 are TO226 package devices, similar to TO92L package, which allows for higher dissipation than acceptable for TO92 type. I believe some revisions used other semis like MPSA93K, as shown in the parts list, with no change in performance or specs.
 
Last edited:
Has anyone ever adapted the Quad 306 floating power supply for other uses? As far as I can see it makes the speaker capacitor coupled at the power supply end. One member of Quads repair staff said he never understood why they did it. Would anyone like to say why it might have been prefered?
 
Well,as i have said i am a newbie with much more ignorance than knowlege...thats the main reason for being here,in order to learn from people who know HOW to try thinks or ideas.

I suspect the better upgrade would be a larger transformer. For safety and simplicity I would go for the largest off the shelf design of slightly lower output voltage. I always thought the 306 sounded a bit too nice. My instinct is the transformer could be larger. In the UK Tiger of Peterborough do reasonably priced custom wound transformers. They might have a larger VA type of similar diameter which is taller, or even one that just fits the box ( never allow the fixing bolt to touch the case at both ends, a shorted turn and overheating ). I would then get a piece of metal cut to allow the transformer to protrude out of the casework ( if larger than the case ). 500VA seems about right. When the idea passes to wanting a better amplifer the reversal to original 306 would be easy. A useful transformer going on to a new project. I have a hunch only this upgrade would really help. One beauty of modifying the 306 this way is that the 306 only has two wires for the transfomer to power supply excluding live and neutral. These two wires have no absolute polarity. Thus very little to get wrong. If buying off the shelf the transformer can have it's windings added in series if that is cheaper. For example 40 Vrms can be 2 x 20 Vrms. Transformers tend to be cheap which is a bonus. Most Naim amps have larger transformers as the main upgrade, certainly the ones of the 306 era.

I would exspect the sound to be more solid and show more detail. Not drastically better. Just more like a 606. The 606 always sounded better even at low volume. All the same the 306 can sound very good. It worked very well with a studio speaker called Orchid PPL1 ( very well is under selling how well ). On paper is couldn't! I never tried it with Klipsche Forte 2. I suspect it would have been excellent, just a hunch.
 
It makes some kind of virtual ground .. so you have no DC coupling to the speaker and reduce the risk to put fire in the voice coil..... and it does not induce low frequency phase shift.

Thank you, that's engineering to see how that could be made to happen. I have wondered about using it as a virtual ground for op amps. Versions I have seen for that have some biasing to the transistors. TLE2426 being a ready made device if 20 mA is enough.
 
Administrator
Joined 2007
Paid Member
It makes some kind of virtual ground .. so you have no DC coupling to the speaker and reduce the risk to put fire in the voice coil..... and it does not induce low frequency phase shift.

That's it exactly :)

The key to understanding how it works is to realise that the centre point of the two reservoir caps is nominated as 'ground' or the reference point to which all voltages are referred back to.

So while the speaker is AC coupled in one sense, the amp also has the ability to maintain a DC voltage across the speaker for a very limited time. That means that the squarewave response at LF will show no roll of as would a 'normal' capacitor coupled load. What does happen though, is that depending on the polarity of the output voltage (and hence current flow in the load) you will see the 'positive' or 'negative' rail fall as the appropriate reservoir cap gives up its charge.

This means that the length of time the output can sustain DC in the load depends on both the load impedance and the amplitude of the signal with both these values working with the time constant of the reservoir cap/s.

Its a really neat and clever system.
 
Member
Joined 2010
Paid Member
I suspect the better upgrade would be a larger transformer....
The 306 transformer is mounted on the PCB, supported by a block of rubber glued to the case beneath and a smaller piece above it. It's a structural weakness of the design that allows damage from vibration in transit.

Upsizing the transformer by any worthwhile amount, would be impractical, I think. If you want a more powerful amp, DIY build one in a larger case - you could use available Quad 606 PCBs if necessary. Just don't spoil the original 306 work of design art :eek:
 
Thanks Ian. I thought as much when going to bed last night. I get so upset when people make changes that can not be reversed. I should have said nothing.

Thanks everyone for going through the virtual ground. Anyone want to say why underbised class B to the two transistors is OK?

Sorry to our question setter for asking about this as it is deep in the circuit. Even the senior service engineer at Quad was baffled by it. He was a very good engineer. It is a most remarkable and cheap idea. I hope one of us tries it in a new design.

One way to do my 306 upgrade is to buy another one. I suspect paralell bridging is OK as with the 405? Paralell bridging will increase current output. It also gives a true Monobloc. It might see off a 606 is so. See the discussion below. Quad did have a service sheet for this when 405. I am tempted to think they would exspect 306 to do the same. I don't know. 405 bridged into Gale 401 speakers was delightful. One 405 really couldn't cope with the 401's.

http://www.diyaudio.com/forums/soli...el-them-use-them-mono-blocks.html#post2140310
 
Last edited:
Administrator
Joined 2007
Paid Member
Thanks everyone for going through the virtual ground. Anyone want to say why under biased class B to the two transistors is OK?

That is the whole essence of current dumping. You have the output stage running in Class C meaning it conducts for less than 180 degrees per device as both are fully cut off under quiescent and low output conditions. Under these conditions the output is supplied from the Class A driver. As the output increases further the main output devices come into play leaving the Class A stage to 'fill in' the discontinuity around the crossover region, the output of the Class A and Class C stages feeding into a 'bridge' to derive a correction signal.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.