Alt-BEAM Archive
Message #03617
To: "'Wouter Brok'" w.j.m.brok@stud.tue.nl,
From: Wilf Rigter Wilf.Rigter@powertech.bc.ca
Date: Mon, 24 May 1999 08:04:42 -0700
Subject: [alt-beam] Re: Suspended master-slawe bicore walker.
Hello Walter,
Going from textbook to reality, one detail you need to consider is that the
switching thresholds are never precisely Vcc/2. In fact with 74HCTxxx
devices, the thresholds are deliberately offset to match TTL logic levels.
In the case of 74HCT240, the Bicore switching occurs relatively early
(at +1.5V) on the rising capacitor charging voltage curve (noise immunity)
and symmetry is achieved by matching the 2 capacitors. This is necessary
to match the input time constants since each stage alternately triggers the
Bicore from the input which has the voltage rising from 0V towards Vcc/2
and which is the always the first to cross the threshold at ~1.5V.
As always there may be something useful in this.
regards
Wilf Rigter mailto:wilf.rigter@powertech.bc.ca
tel: (604)590-7493
fax: (604)590-3411
> -----Original Message-----
> From: Wouter Brok [SMTP:w.j.m.brok@stud.tue.nl]
> Sent: Monday, May 24, 1999 1:34 AM
> To: Jakub Pietracha
> Cc: beam@corp.sgi.com
> Subject: Re: Suspended master-slawe bicore walker.
>
> Hello Jakub,
>
> No, as you describe it it isn't a suspended bicore anymore. Actually it's
> a
> normal bicore (i.e. a ringlike structure of two Nv-neurons) but with one
> restiction: the resistors (of the two neurons) are dependent on eachother
> in such a way that the total resistance will be constant (if one
> resistance
> is made bigger by turning the pot the other one is made smaller by the
> same
> amount).
>
> The resulting behaviour from this replacement is quite different from the
> behaviour of the original master suspended bicore. For example, now it
> will
> be possible to control the duty-cycle in a better way (and thus the swing
> of the legs).
>
> By saying this I realise that I suggest that the suspended bicore has a
> dutycycle, different from 50%, which is harder to control then in the new
> situation. This might raise some questions, because I think most people
> believe that the duty-cycle of the suspended bicore is 50% by defenition.
> This however is not generally true, although one would expect the average
> dutycycle to be 50%. The duty-cycle of on period of oscillation does not
> need to be 50%. Actually it can differ quite a bit from that because of
> noise.
>
> For a suspended bicore with equal capacitors and inverters which have
> threshold-voltages approximately Vcc/2, the moment of switching states is
> very likely to be caused by noise. And if this is so the circuit can
> switch
> state in the first half a period, for example, earlier than in the second
> half of the period, which results in a duty-cycle for that period,
> different from 50%.
> One would expect that the average duty-cycle is 50% since noise is
> (thought
> of to be) stochastic and doesn't favour one particular part of the period
> of oscillation. However if it does favour one paricular part of the period
> the resulting average duty-cylce differs from 50%. Is this possible I
> wonder ... Yes, I think it is (although perhaps not the reason for the
> nonsymmetric leg-swing in your question) because maybe the noise produced
> by the motor is different in amplitude for the two different directions.
> If
> the noise influences the master suspended bicore this then will result in
> an average duty-cycle not equal to 50%.
>
> A couple of weeks back somebody posted the same question to the list ...
> same problem. Maybe what I described above is the reason of the problem;
> any comments from other people on the list?
>
> Jakub, the solution you choose is a good one to really control the
> duty-cycle, but I don't know if it is what you want. Like this the circuit
> is less sensitive to noise and maybe you actually wanted it to be
> sensitive. If so I would suggest that you connect springs to the legs, so
> that they will not turn all the way around eventually and accept the fact
> the the walker doesn't walk in a straight line.
> If what I described is the reason (maybe you will get more reactions later
> proving me wrong (which I would like to know)) then I you could make the
> circuit less sensitive to noise by using capacitors decoupling
> power-supply-lines and by choosing different-valued capacitors in the
> master suspended bicore.
>
> Come to think of it ... it is the moter connected to the master which
> turns
> non-symmetrically is it?
>
>
> Regards,
>
> Wouter Brok.
>
>
>
------------------------------------------------------------------------
eGroups.com home: http://www.egroups.com/group/alt-beam
http://www.egroups.com
- Simplifying group communications
Home