BELL & HOWELL: Model 621 projector amplifier, 1951

Despite saying never again - I now have another Bell & Howell (Filmosound) amp on the bench.   This is a 621 model from 1951, which as far as the amp is concerned, looks nearly identical to my 601 from 1949.  I believe the main differences were with the actual projector itself - the 621 being a later and improved model.

Unlike the last amp - where I rebuilt the amplifier on tag boards following the original schematic—I'm aiming to retain as much of the existing amp as I can this time. It’s slow going, but I’ve got a better feel for it after the last project. That said, plans may still change.

Rear panel - unusual power connector to be removed and replaced with an IEC.  Odd speaker connectors will also go.  I could leave it original, but who wants to lose a weird connector and be left with an amp that doesn't work?

The valve on the far left will end up in the bin - not an audio valve.  The rest look pretty nice.  They have 3 x 6V6, only two are used for audio, so I have a spare.  Being a cathode biased amp, matching is not critical.

Before I got started - how about we just call it "an efficient use of space" and leave it at that.  This is why a lot of people just gut them and start fresh.  You literally can't fit a soldering iron down in some areas.

Clean-out underway – removing all non-audio components and doing a bit of rewiring to make things easier to work on.

It came with a friend (on the right) – a widow-maker type without a power transformer. Useful for parts, but not a viable option for a guitar amp.

BELL & HOWELL MODEL 621 SCHEMATIC & LAYOUT

Full user manual can be found here.

Probably the most interesting part of the amp is the tone control.  It's quite different to the better known model 385 Filmosound.  Not even a close comparison.

Instead of the familiar FMV or James/Bax stack that sits between two stages, this amp has its entire tone circuit right on the cathode of the ECC35 triode.  A single 50 kΩ pot and three capacitors juggle the amount of local feedback inside that valve and the amount of global negative feedback (NFB) returned from the output transformer. The result is a surprisingly flexible, low-loss, one-knob tone control.

MODIFIED SCHEMATIC

Trimmed back to the essentials.  I removed some of the weird feedback around the preamp pentode, as it doesn't really add anything useful.

I tried to match the numbering to the original schematic as much as possible.

My notes are pretty bad - but you can kind of see what I chopped out.

TONE STACK: James

The James is one of the more common Bass / Treble controls, most often found in hi-fi, or amps that aren't Fender or Marshall...  although they did use them on rare occasions.  Perhaps the most well known example is in Orange amplifiers and some Ampegs.   It was used in the Marshall Artiste, and a blonde Fender Twin (6G8).

It's hard to mention James without Baxandall - same same but different, so it will not be included in this post.  The original Bax is an active control designed for hi-fi - it is incorporated into a gain stage.  The James predates the Bax by a few years and is a passive circuit (1949 for James vs 1952 for Bax).  

It's a simple design that leaves the mids mostly alone and the bass and treble controls work independently of one another (for the most part).  Think of it as two tone control operating in parallel.  Ignore RIN and RL - these are the loads of the stage either side of the tone controls.   

ORANGE, GRAPHIC MKII

link to TSC for frequency response

AMPEG B15

FENDER TWIN, 6G8

note: this uses a tapped pot.  i.e. it has fourth lug - good luck finding one of those.

MARSHALL ARTISTE

VOX: AC2 & AC30 Schematics, EL34 based

 The little known EL34 Vox amps from the late fifties, one of which is fixed bias.   https://www.voxac30.org.uk/vox_ac2_30_1957.html. 

VOX AC2, 30 WATT AMPLIFIER.  OS/003 (FIXED BIAS)

VOX AC30, NO2.  OS/007  1960 (CATHODE BIASED)

EL34: individual 470 ohm cathode resistor and 50uf bypass caps.  5k output transformer

VOX: Vintage AC10 / AC15 Amplifier Schematics

The vintage Vox AC10 & AC15 all have the same phase inverter and output stage, with some very minor variations.   Popping these in as a reference for myself - yes, I have simplified some things / made generalisations.   If you're interested in ACs in general - https://www.voxac30.org.uk/. 

POWER SUPPLY

• Power transformer secondaries: 300-0-300, 120mA
• Rectifier:  EZ81, 320V DC output
• Filtering:  16uf - 8H - 16uf (mostly) 
• Phase inverter:  usually 22k, 8uf.  Sometimes 47k
• Preamp:  As above

POWER AMP / OT

• 8k primary, 8 & 15ohm secondaries 
• EL84 Plates: ~305 - 310V
• Cathode bias resistor: 130ohms, 5-8w
• Cathode bypass: 50uf
• Plate and screen resistors:  either one or the other, never both - 100 ohm
• Grid stopper: Usually 1k5, 10k on the 1959 AC15
• Grid leak: 220k
• Coupling caps from phase inverter: 47n/50n on AC10.  Mix of 50n & 10n on AC15
• Cut:  5n & A250k (10n on AC15 1960)

Expect about 10v on the cathode. 

PHASE INVERTER

Vintage Vox amps have two channels, each using a side of the phase inverter as an input.  For a single channel amp, the unused input of the phase inverter must be grounded with a capacitor - 47n - 100n.

• Classic long-tail pair, ECC83
• 100k plate resistors
• Coupling caps to EL84s: 47n/50n on AC10.  Mix of 50n & 10n on AC15
• 1meg, 1k2, 47k combo for grids / cathode (sometimes 1k)
• Coupling cap on input: 10n & 100n
• Volume pots from previous stage: A500k

VOX AC/10 OS/008 1960

VOX AC-10  NO. V-1-3

VOX AC/15  NO. 2  OS/005 1959

VOX AC15  NO. 3  OA/031 1960

VOX AC-15  NO. V-1-5 

UNKNOWN: Dutch EL84 Valve PA head

Unknown amp from the Netherlands….   I really have no idea about this one, and I’ve had no luck searching online.  I can only assume that it's from the Netherlands based on the control labels being Dutch.

I was drawn to it mostly for the look, and the fact that it had 2 x EL84s and an output transformer with regular speaker impedances on the secondaries (and the price was OK).  Apart from these factors, the tag strip running straight down the middle of the chassis was appealing - looked like it might be handy for mods or rebuilding.  Beyond that, the internals are… well, kind of a mess. It looks like someone has been in here before and done some questionable work.

Anyway - despite looking a mess with questionable work, it actually works really well.  I wasn't sure if it would even work, but it sounds great, and has zero hum.   Zero hum, and not a single heater wire is twisted, nor has any effort been made to keep it away from other components / wires. 

ARCHITECTURE

Input jack → EF86 preamp → Volume → 12AX7 gain stage → James‐style bass/treble tone stack → 12AX7 gain stage → Master Volume → 12AX7 floating paraphase inverter → EL84 push-pull output pair.   

The phono input basically skips the EF86 stage and the rest is the same as above - straight into a volume control and forward from there.

Negative feedback from the output transformer, back to the gain stage of the phase inverter.

INPUTS

• mic
• pick-up (phono)

CONTROL PANEL 

• Microfoon
• Pick-up
• Toon
• Toon
• Master volume / power on-off

VALVES

• Tesla EF86 preamp
• 2 x Mullard 12AX7 - phase inverter and gain stages for the tone controls
• 2 x EL84
• EZ80 rectifier, labelled Made in Holland    

OUTPUT TRANSFORMER

• H.STOET'S RADIO
• Primary  7-8k
• Secondaries:  4, 7 and 15 ohm
• 11 watts
• 30hz - 20khz

POWER TRANSFORMER

• Primaries:  120v & 220v
• Secondaries:  270, 0, 270 and 6.3 heaters and an unused pair for 5v   
• 100mA

CHOKE

• 6H
• 100mA
• 150ohms

B+ 

• ~300V @ 220V

 

SCHEMATICS

JAMES EQ RESPONSE (UNMODDED)

Response taken after the 10n cap, before the master volume.  I quite like the sound of it, but I did give it a slight tweak to shift the centre point lower, and reduce the amount of scoop.

MODS

I didn't really have to do anything major circuit wise - just a few tweaks here and there.

• removed the phono input

• moved the input jack from the back to the front

• speaker jack added on the rear

• It still needs a fuse on the AC input, but I'll get around to that shortly.  Probably a new power lead at the same time.

• Removed low value cathode bypass caps

• Adjusted the James EQ a little

• Removed negative feedback

REPAIRS

I did find a couple of things that needed sorting out, some of which changed the tone quite a bit.  

• Bad 100k resistor on the plate of the EF86 and a large 33k dropping resistor in the filter section for the same - I was only getting 35volts on the plate of the EF86.  Mind you, it did sound kind of cool and dirty.

• 2k2 on the cathode of the 12AX7 triode after the James EQ gain stage was creating more noise than I was happy with (of the rushing variety).

• There was a 100p cap from C5 to ground, which I'm assuming may have been put there to try and reduce noise (possibly from the dodgy cathode resistor)

• Random small value cathode bypass caps removed from the 12AX7s either side of the James EQ (6n6 & 4n7) - have no idea why they were even there.  Small values only boost highs.

HEATER VOLTAGE

This amp was originally made for 220V mains, but here in Australia we get ~240V (even though it's technically 230V).  While it's not a big deal for the B+ that small increase can push your heater voltage higher than it should be.  The transformer already ran a bit hot, putting out 6.45V on 220V, while well within the 6.3V +/- 10% specification, it didn't help my situation at all.  Running at 240v it was reading well over 7V and that’s not going to be good for valve life.

To fix this, I added two 1N5408 diodes in series with one side of the heater winding, right after the power transformer.  These drop about 0.7V each, bringing the heater voltage back into a happy place.  The 1N5408s are rated for 3A at 1000V, so they can easily handle the heater current.  Other diodes could work too, but these did the job perfectly in this case.