When I was a teenager I became interested in a sound I heard on a lot of records from the 60's popularised by groups such as the Animals and the Doors - the rock organ. With the advance of synthesizer technology in the early 80's dedicated rock organs soon began to disappear. Two last attempts were made on the market by Roland with their VK-09, and Korg with their BX-3 and CX-3. All of these instruments tried to produce the sound of a Hammond tonewheel organ using solid state electronics. By the mid eighties there was nothing on the market and I started to investigate building my own, then I managed to obtain a VK-09. It worked, but in truth it's organ sounds were no better than those of a decent synth, and certainly no substitute for a Hammond.
The Vox continental used by the Animals amongst others was an early attempt to produce the sound of a Hammond in an easily portable transistorised unit. The unique sound of a Hammond organ comes from its mechanical tone generation system. A series of toothed wheels are rotated at constant speed by motor locked to the incoming mains frequency. Each wheel has an electromagnetic pickup mounted near it. The wheel is carefully shaped such that the output from the pickup is a (nearly) pure sine wave. These signals are then switched in and out by the keyboard and mixed in varying proportions using the 9 drawbars. The result is then amplified and fed to a loudspeaker, possibly a rotating Leslie unit. The drawbars allow for infinite variation of the sound in a way that a simple fixed pre-set on a synth cannot match. Less often mentioned is the way they can also compensate for room acoustics: too bassy push in the 16', 5 3/5' & 8', too muffled pull out the 1'.
During the 90's something of a Hammond revival was taking place, especially sought-after were the valve based B3 models and relatives whose value had shot through the roof. Suzuki in Japan revived the Hammond brand and produced a modern range of instruments dedicated to reproducing the sound of the B3 series. Roland have also re-entered the market with their VK-07 model.
A couple of years back I obtained a Leslie rotating speaker module that had been ripped out of an organ. Playing the VK-09 through this was certainly an improvement, but my Ensoniq SQ1 Synth was still more convenient. Knowing of my interest in such things a friend told me of an advert in his local paper for something called a Hammond T-500. I went to see it, fell in love with the sound and re-arranged my furniture to fit it in:
Compared to a B3 the T-500 is about half the weight, solid state, has built in rhythm unit and Leslie, and can be obtained more cheaply. On the downside it also has fewer tonewheels meaning no 16' drawbar on the lower manual. It also has no 'wrap-around', towards the top end of the keyboard. The generators cut out instead of repeating as in a B3. This means the characteristic Hammond soloing wail is absent.
The tone generators fitted to these later organs seem more mellow sounding than the earlier units, presumably the sinewave produced is purer, but this can make it sound dull with the 88800000 style patches often quoted in Hammond literature.
Status: Working
Restoration Problems: Various bad connections, loose key.
The XM-1 has very good sound, better than T-500 on church tones and classic 88800000 style patches, but you can't beat the feel of the real thing (and the smell of hot oil after an hour or so!). The nine drawbars on the controller are also shared between Upper, lower and pedals. This makes it more difficult to make instant corrections since the slider's positions may physically match the Upper manual, whilst you are needing to adjust the lower (or visa versa).
Whilst Guitarists are at least as concerned about their amplifiers as their guitars, there doesn't seem to be much interest in keyboard amplification outside of the Leslie. Modern keyboards can sound rather sterile and clean - running them through a valve (tube) amplifier can inject the warmth that's missing. If the amp is driven gently into distortion then a nice 'creamy' 'gimme some loving' sound results. At least part of the classic Leslie sound is due to its relatively low power amp distorting.
I started by obtaining a cheap secondhand guitar head hoping that I could modify it for keyboard use. A guitar amplifier is fundamentally a midrange device, whilst a modern keyboard needs a wide frequency range, especially if a drum machine is used. On trying to improve the performance of the guitar amp by changing the turnover frequencies of the tone controls and increasing the feedback around the power amplifier two things quickly became clear:
1 The output transformer was poor in terms of both
primary and leakage inductance.
2 The output tranformer was limiting the power from
2 EL34's to about 35watt.
3 The typical cascading of ECC83s in the preamp stages
leads to a poor top end response due to Miller effect.
At this point I gave up and started to look for something more suitable. In fact the requirements of a modern keyboard amplifier are more akin to a good PA amplifier. I was lucky to pick up a Sound City ST50 PA amp from the late sixties/early seventies. This also uses 2 EL34's in push-pull but is fitted with Partridge transformers. Partridge were the premier English transformer maker in the valve era and it shows. The performance of the power amplifier was excellent, delivering over 50Watts at low distortion, however the pre-amp section left much to be desired and there was plenty of hum.
The majority of the hum was eventually traced to the power amp, unusually this amplifier uses a concertina type phase splitter. A well-known disadvantage of this circuit is that the Cathode and Anode impedances do not match. Because the fixed bias for the output valves is applied at these points different amounts of hum are being injected into the two sides of the output stage. The means that cancellation in the output transformer is not very effective and results in large amounts of audible hum. The cure was to increase the bias smoothing capacitor and add a second stage of filtering.
At this point I had a nice quiet power amplifier albeit with a rather large intput voltage requirement of 4V RMS for 50Watt out. The pre-amplifier is split into 2 channels of seperate tone control stages, each with 3 inputs, 2 of the inputs are high sensitivity and one low. The low sensitivity input needed approx 2V for full output - this is a pretty high figure. The high sensitivity inputs needed about 50mV a bit low for a microphone, but a bit high for keyboard or other 'line' level sources.
The tone controls were not proper Baxandall type and were being driven by an ECC83 with 15V RMS on its anodes. This is a pretty large output and measured distortion was very poor. The tone controls range was clearly compromised by the need to keep the drive level from the previous stage down to a 'sensible' level. The mixing of the 3 inputs before the tone stage was by simple resistors, leading to large amounts of interaction between controls (adjusting the level of one input also affects the others).
There seemed no alternative but to re-design the pre-amplifier and build in the necessary versatility, whilst making the inputs more sensibly rated. In the end I added 2 extra EF86 valves used as a virtual earth mixers and driving a proper Baxandall passive tone circuits designed using Spice simulation. The output from the 2 tone stages was then amplified and combined in a final ECC82. In this way no excessively large swings were needed in the pre-amp. The gain of the EF86 stage was fixed by feedback such that the low sensitivity inputs were now a much more usable 1V. I then changed and re-biased the input valves for the high sensitivity iputs from ECC81's to ECC83's. This gave a max sensitivity of about 15mV for full output I also added a switchable pad so that these inputs could also take line inputs of about 150mV and up.
The new valves were installed on L shaped aluminium brackets attached to the front panel and careful attention paid to eathing and shielding (especially from the 2 neon power indicators!). The modified unit has been in use for about 5 years now and continues to perform very nicely. It is normally used together with a Torque Speaker.
Click here for more ST50/PA120 photos
I have recently obtained a Sound City PA120 which uses 6 EL34's for an output of 120Watts. It has 6 inputs, each with seperate bass and treble controls. However technically this amplifier also has its problems and I am incorporating some of the ideas from the ST50 in a re-design.
In the case of the PA120 it uses the same power amp as the Sound City 120 series guitar amps, complete with capacitor from the feedback line to ground. This means that achieving stabilty was easy for them, but the power amp section will have a rising frequency response. This means that if the 'reverb in' is used as the input, so the power amp effectively becomes a slave, the results will be pretty unbalanced (also true if 'reverb out' were used to drive another power amp). I wanted to have the ability to slave the 120 off of the ST50 so wanted a flat frequency response. Also tapping off feedback like this will lead to high levels of distortion at high frequencies.
In its original state the background hum and buzz levels were also poor.
What I have done is to move the feedback line from the phase splitter to the input stage and add compensation as necessary. This gives nice flat response on 4 & 16 ohm taps (although some transformer winding issue seems to make the performance not quite as good on 8ohm). To keep the feedback -ve obviously the phase splitter outputs have to be exchanged.
Signal to noise ratio (power amp) is now ~87dB which is an incredible result for a PA amp.
The resulting circuit now bears an uncanny resemblance to the Hiwatt PAs, although I wasn't aware of this at the time - great minds think alike?? :-)
The rising response of the power amp meant that the tone controls of the preamp are a bit wierd (because the treble boost is coming from the power section). This is OK as far as it goes, but Spice simulation and measurement of the real thing showed that a flat response was not really on offer. I have designed new tone controls (similar to what I put in my ST50) and then changed the first valve to a 12AT7 which is better at driving the load of a typical tone control circuit.
To allow the use of the 'reverb out' socket to drive other gear and long cables without high frequency loss I have added another valve. A 6C4 (1/2 ECC82) is used as a cathode follower after the mixer summing resistors. This is also of benefit for driving the internal screened cables.
Despite the model number nearly 150watts are on offer, which makes sense
given that in parts terms it's effectively 3*the 50 watter.
